SlideShare a Scribd company logo

Clustered data ontap_8.3_physical_storage

M
Mohan Sati

This document discusses physical storage management for NetApp's Clustered Data ONTAP operating system. It provides information on managing disks and array LUNs, including disk types, speeds, and formats. It also covers topics like root-data partitioning, disk ownership, spare management, encryption, and array LUN configuration.

Technology
1 of 212
Download to read offline
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
Clustered Data ONTAP® 8.3 Physical Storage Management Guide NetApp, Inc. 495 East Java Drive Sunnyvale, CA 94089 U.S. Telephone: +1 (408) 822-6000 Fax: +1 (408) 822-4501 Support telephone: +1 (888) 463-8277 Web: www.netapp.com Feedback: doccomments@netapp.com Part number: 215-09163_B0 March 2015
Clustered data ontap_8.3_physical_storage
Contents Managing disks using Data ONTAP ......................................................... 10 How Data ONTAP reports disk types ....................................................................... 10 Storage connection types and topologies supported by Data ONTAP ...................... 12 How disks can be combined for the SAS storage connection type ............... 12 How disks can be combined for the FC-AL storage connection type ........... 12 Methods of calculating aggregate and system capacity ............................................ 13 Disk speeds supported by Data ONTAP ................................................................... 13 How drive checksum types affect aggregate and spare management ....................... 14 Drive name formats ................................................................................................... 14 Pre-cluster drive name formats ................................................................................. 15 Loop IDs for FC-AL connected disks ........................................................... 18 Understanding RAID drive types .............................................................................. 18 How disk sanitization works ..................................................................................... 19 Disk sanitization process ............................................................................... 19 When disk sanitization cannot be performed ................................................ 20 What happens if disk sanitization is interrupted ........................................... 20 Tips for creating and backing up aggregates containing data to be sanitized ................................................................................................... 21 How Data ONTAP monitors disk performance and health ....................................... 21 What happens when Data ONTAP takes disks offline ................................. 21 How Data ONTAP reduces disk failures using Rapid RAID Recovery ....... 21 How the maintenance center helps prevent drive errors ............................... 22 When Data ONTAP can put a disk into the maintenance center .................. 23 How Data ONTAP uses continuous media scrubbing to prevent media errors ........................................................................................................ 24 How you can use ACP to increase storage availability for SAS-connected disk shelves ................................................................................................................. 24 How you use SSDs to increase storage performance ................................................ 25 How the All-Flash Optimized personality affects node behavior ................. 25 How Data ONTAP manages SSD wear life .................................................. 26 Capability differences between SSDs and HDDs ......................................... 26 Guidelines and requirements for using multi-disk carrier storage shelves ............... 27 Table of Contents | 3
How Data ONTAP avoids RAID impact when a multi-disk carrier must be removed .............................................................................................. 27 How to determine when it is safe to remove a multi-disk carrier ................. 28 Spare requirements for multi-disk carrier disks ............................................ 28 Shelf configuration requirements for multi-disk carrier storage shelves ...... 29 Aggregate requirements for disks in multi-disk carrier storage shelves ....... 29 Considerations for using disks from a multi-disk carrier storage shelf in an aggregate ............................................................................................. 29 Understanding root-data partitioning ........................................................................ 30 How root-data partitioning works ................................................................. 30 How root-data partitioning affects storage management .............................. 31 Which drives are partitioned and used for the root aggregate ....................... 32 Standard root-data partitioning layouts ......................................................... 32 Requirements for using root-data partitioning .............................................. 34 Initializing a node to configure root-data partitioning .................................. 34 Setting up an active-passive configuration on nodes using root-data partitioning .............................................................................................. 39 Adding disks to a node .............................................................................................. 41 When you need to update the Disk Qualification Package ........................... 42 Replacing disks that are currently being used in an aggregate ................................. 43 Replacing a self-encrypting disk ............................................................................... 44 Converting a data disk to a hot spare ........................................................................ 44 Removing disks from a node ..................................................................................... 45 Removing a failed disk .................................................................................. 45 Removing a hot spare disk ............................................................................ 46 Removing a data disk .................................................................................... 47 Using disk sanitization to remove data from disks ................................................... 48 Stopping disk sanitization ......................................................................................... 51 Commands for managing disks ................................................................................. 51 Commands for displaying information about storage shelves .................................. 52 Commands for displaying space usage information ................................................. 53 Managing ownership for disks .................................................................. 54 Types of disk ownership ........................................................................................... 54 Reasons to assign ownership of disks and array LUNs ............................................ 54 How disks and array LUNs become available for use .............................................. 55 How automatic ownership assignment works for disks ............................................ 56 4 | Physical Storage Management Guide
Which disk autoassignment policy to use ..................................................... 56 When automatic ownership assignment is invoked ...................................... 57 How disk ownership works for platforms based on Data ONTAP-v technology ..... 57 Guidelines for assigning ownership for disks ........................................................... 58 Assigning ownership for disks .................................................................................. 58 Assigning ownership for disks partitioned for root-data partitioning ....................... 59 Removing ownership from a disk ............................................................................. 60 Configuring automatic ownership assignment of disks ............................................ 61 How you use the wildcard character with the disk ownership commands ................ 62 Managing array LUNs using Data ONTAP ............................................. 64 Data ONTAP systems that can use array LUNs on storage arrays ........................... 64 Installing the license for using array LUNs .............................................................. 65 How ownership for disks and array LUNs works ..................................................... 66 Reasons to assign ownership of disks and array LUNs ................................ 66 How disks and array LUNs become available for use .................................. 66 What it means for Data ONTAP to own an array LUN ................................ 68 Why you might assign array LUN ownership after installation .................... 68 Examples showing when Data ONTAP can use array LUNs ....................... 69 Assigning ownership of array LUNs ......................................................................... 71 Verifying the back-end configuration ....................................................................... 72 Modifying assignment of spare array LUNs ............................................................. 73 Array LUN name format ........................................................................................... 74 Pre-cluster array LUN name format .......................................................................... 74 Checking the checksum type of spare array LUNs ................................................... 76 Changing the checksum type of an array LUN ......................................................... 76 Prerequisites to reconfiguring an array LUN on the storage array ........................... 77 Changing array LUN size or composition ................................................................. 78 Removing one array LUN from use by Data ONTAP .............................................. 79 Preparing array LUNs before removing a Data ONTAP system from service ......... 79 Securing data at rest with Storage Encryption ........................................ 81 Introduction to Storage Encryption ........................................................................... 81 What Storage Encryption is ........................................................................... 81 How Storage Encryption works .................................................................... 82 Disk operations with SEDs ........................................................................... 82 Benefits of using Storage Encryption ............................................................ 83 Limitations of Storage Encryption ................................................................ 84 Table of Contents | 5
Setting up Storage Encryption ................................................................................... 85 Information to collect before configuring Storage Encryption ..................... 85 Using SSL for secure key management communication .............................. 86 Running the Storage Encryption setup wizard .............................................. 88 Managing Storage Encryption ................................................................................... 90 Adding key management servers .................................................................. 90 Verifying key management server links ........................................................ 91 Displaying key management server information .......................................... 92 Removing key management servers .............................................................. 93 What happens when key management servers are not reachable during the boot process ....................................................................................... 94 Displaying Storage Encryption disk information .......................................... 95 Changing the authentication key ................................................................... 96 Retrieving authentication keys ...................................................................... 97 Deleting an authentication key ...................................................................... 98 SSL issues due to expired certificates ........................................................... 98 Returning SEDs to unprotected mode ......................................................... 100 Destroying data on disks using Storage Encryption ................................................ 102 Sanitizing disks using Storage Encryption before return to vendor ............ 102 Setting the state of disks using Storage Encryption to end-of-life .............. 103 Emergency shredding of data on disks using Storage Encryption .............. 104 What function the physical secure ID has for SEDs ................................... 105 SEDs that have PSID functionality ............................................................. 106 Resetting an SED to factory original settings ............................................. 106 How Data ONTAP uses RAID to protect your data and data availability ............................................................................................ 108 RAID protection levels for disks ............................................................................. 108 What RAID-DP protection is ...................................................................... 108 What RAID4 protection is ........................................................................... 109 RAID protection for array LUNs ............................................................................ 109 RAID protection for Data ONTAP-v storage ......................................................... 110 Protection provided by RAID and SyncMirror ....................................................... 110 Understanding RAID drive types ............................................................................ 113 How RAID groups work ......................................................................................... 113 How RAID groups are named ..................................................................... 114 Considerations for sizing RAID groups ...................................................... 114 6 | Physical Storage Management Guide
Customizing the size of your RAID groups ................................................ 115 Considerations for Data ONTAP RAID groups for array LUNs ................ 116 How Data ONTAP works with hot spare disks ....................................................... 117 Minimum number of hot spares you should have ....................................... 117 What disks can be used as hot spares .......................................................... 118 What a matching spare is ............................................................................ 118 What an appropriate hot spare is ................................................................. 118 About degraded mode ................................................................................. 119 How low spare warnings can help you manage your spare drives .............. 120 How Data ONTAP handles a failed disk with a hot spare ...................................... 120 How Data ONTAP handles a failed disk that has no available hot spare ............... 121 Considerations for changing the timeout RAID option .......................................... 121 How RAID-level disk scrubs verify data integrity .................................................. 122 Changing the schedule for automatic RAID-level scrubs ........................... 122 How you run a manual RAID-level scrub ................................................... 123 Controlling the impact of RAID operations on system performance ...................... 123 Controlling the performance impact of RAID data reconstruction ............. 124 Controlling the performance impact of RAID-level scrubbing .................. 124 Controlling the performance impact of plex resynchronization .................. 125 Controlling the performance impact of mirror verification ........................ 126 What aggregates are ................................................................................. 127 How unmirrored aggregates work ........................................................................... 127 How mirrored aggregates work ............................................................................... 129 What a Flash Pool aggregate is ............................................................................... 130 How Flash Pool aggregates work ................................................................ 130 Requirements for using Flash Pool aggregates ........................................... 131 Considerations for RAID type and spare management for Flash Pool cache ...................................................................................................... 132 How Flash Pool aggregates and Flash Cache compare ............................... 133 How the available Flash Pool cache capacity is calculated ........................ 133 Using caching policies on volumes in Flash Pool aggregates ..................... 135 Modifying caching policies ......................................................................... 136 How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates .................................................................................................. 137 Considerations for when to use SSD storage pools ..................................... 138 How you use SSD storage pools ................................................................. 139 Table of Contents | 7
Requirements and best practices for using SSD storage pools ................... 139 Creating an SSD storage pool ..................................................................... 140 Adding SSDs to an SSD storage pool ......................................................... 141 Considerations for adding SSDs to an existing storage pool versus creating a new one ................................................................................. 142 Determining the impact to cache size of adding SSDs to an SSD storage pool ........................................................................................................ 143 Commands for managing SSD storage pools .............................................. 143 How the SVM affects which aggregates can be associated with a FlexVol volume ............................................................................................................... 144 Understanding how Data ONTAP works with heterogeneous storage ................... 145 How you can use disks with mixed speeds in the same aggregate ............. 145 How to control disk selection from heterogeneous storage ........................ 145 Rules for mixing HDD types in aggregates ................................................ 146 Rules for mixing drive types in Flash Pool aggregates ............................... 147 Rules for mixing storage in array LUN aggregates ..................................... 147 How the checksum type is determined for array LUN aggregates ......................... 148 How to determine space usage in an aggregate ....................................................... 148 How you can determine and control a volume's space usage in the aggregate ....... 150 How Infinite Volumes use aggregates .................................................................... 152 Aggregate requirements for Infinite Volumes ............................................ 152 How FlexVol volumes and Infinite Volumes share aggregates .................. 152 How storage classes affect which aggregates can be associated with Infinite Volumes .................................................................................... 153 How aggregates and nodes are associated with Infinite Volumes .............. 154 How space is allocated inside a new Infinite Volume ................................ 155 Relocating ownership of aggregates used by Infinite Volumes .................. 156 Managing aggregates ............................................................................... 159 Creating an aggregate using unpartitioned drives ................................................... 159 Creating an aggregate using root-data partitioning ................................................. 161 Increasing the size of an aggregate that uses physical drives ................................. 163 Increasing the size of an aggregate that uses root-data partitioning ....................... 165 Correcting misaligned spare partitions .................................................................... 169 What happens when you add storage to an aggregate ............................................. 170 Creating a Flash Pool aggregate using physical SSDs ............................................ 171 Creating a Flash Pool aggregate using SSD storage pools ...................................... 173 8 | Physical Storage Management Guide
Determining Flash Pool candidacy and optimal cache size .................................... 175 Determining and enabling volume write-caching eligibility for Flash Pool aggregates .......................................................................................................... 178 Changing the RAID type of RAID groups in a Flash Pool aggregate .................... 180 Determining drive and RAID group information for an aggregate ......................... 181 Relocating aggregate ownership within an HA pair ............................................... 182 How aggregate relocation works ................................................................. 183 How root-data partitioning affects aggregate relocation ............................. 184 Relocating aggregate ownership ................................................................. 184 Commands for aggregate relocation ........................................................... 187 Key parameters of the storage aggregate relocation start command ........... 187 Veto and destination checks during aggregate relocation ........................... 188 Assigning aggregates to SVMs ............................................................................... 190 Methods to create space in an aggregate ................................................................. 191 Determining which volumes reside on an aggregate .............................................. 192 Determining whether a Flash Pool aggregate is using an SSD storage pool .......... 193 Commands for managing aggregates ...................................................................... 193 Storage limits ............................................................................................ 195 Copyright information ............................................................................. 197 Trademark information ........................................................................... 198 How to send comments about documentation and receive update notification ............................................................................................ 199 Index ........................................................................................................... 200 Table of Contents | 9
Managing disks using Data ONTAP Disks (sometimes also called drives) provide the basic unit of storage for storage systems running Data ONTAP that use native storage shelves. Understanding how Data ONTAP uses and classifies disks will help you manage your storage more effectively. How Data ONTAP reports disk types Data ONTAP associates a type with every disk. Data ONTAP reports some disk types differently than the industry standards; you should understand how Data ONTAP disk types map to industry standards to avoid confusion. When Data ONTAP documentation refers to a disk type, it is the type used by Data ONTAP unless otherwise specified. RAID disk types denote the role a specific disk plays for RAID. RAID disk types are not related to Data ONTAP disk types. For a specific configuration, the disk types supported depend on the storage system model, the shelf type, and the I/O modules installed in the system. The following tables show how Data ONTAP disk types map to industry standard disk types for the SAS and FC storage connection types, storage arrays, and for virtual storage (Data ONTAP-v). SAS-connected storage Data ONTAP disk type Disk class Industry standard disk type Description BSAS Capacity SATA Bridged SAS-SATA disks with added hardware to enable them to be plugged into a SAS-connected storage shelf. FSAS Capacity NL-SAS Near Line SAS MSATA Capacity SATA SATA disk in multi-disk carrier storage shelf SAS Performance SAS Serial-Attached SCSI SSD Ultra-performance SSD Solid-state drives 10 | Physical Storage Management Guide
FC-connected storage Data ONTAP disk type Disk class Industry standard disk type Description ATA Capacity SATA FCAL Performance FC Storage arrays Data ONTAP disk type Disk class Industry standard disk type Description LUN N/A LUN A logical storage device backed by storage arrays and used by Data ONTAP as a disk. These LUNs are referred to as array LUNs to distinguish them from the LUNs that Data ONTAP serves to clients. Virtual storage (Data ONTAP-v) Data ONTAP disk type Disk class Industry standard disk type Description VMDISK N/A VMDK Virtual disks that are formatted and managed by VMware ESX. Related concepts Rules for mixing HDD types in aggregates on page 146 Storage connection types and topologies supported by Data ONTAP on page 12 Related references Understanding RAID drive types on page 18 Managing disks using Data ONTAP | 11
Related information NetApp Technical Report 3437: Storage Subsystem Resiliency Guide NetApp Hardware Universe Storage connection types and topologies supported by Data ONTAP Data ONTAP supports two storage connection types: Serial-Attached SCSI (SAS) and Fibre Channel (FC). The FC connection type supports three topologies: arbitrated loop, switched, and point-to- point. • SAS, BSAS, FSAS, SSD, and MSATA disks use the SAS connection type. SAS-connected storage shelves are connected to the controller on a daisy chain called a stack. • FC and ATA disks use the FC connection type with an arbitrated-loop topology (FC-AL). FC-connected storage shelves are connected to the controller on a loop. • Array LUNs use the FC connection type, with either the point-to-point or switched topology. You cannot combine different connection types in the same loop or stack. However, for MetroCluster configurations, the FC and SAS connection types can be combined in a bridged connection, with FC on the controller side and SAS on the shelf side. The bridged connection can be used in either a direct-attached or switched topology. For more information, see Configuring a MetroCluster system with SAS disk shelves and FibreBridge 6500N bridges in 7-Mode. How disks can be combined for the SAS storage connection type You can combine SAS-connected storage shelves containing performance disks and SAS-connected storage shelves containing capacity disks within the same stack, although this configuration is not recommended. Each SAS-connected storage shelf can contain only one class of disk (capacity, performance, or SSDs). The only exception to this rule is if the shelf is being used for a Flash Pool aggregate. In that case, for some SSD sizes and shelf models, you can combine SSDs and HDDs in the same shelf. For more information, see the Hardware Universe. How disks can be combined for the FC-AL storage connection type You cannot combine storage shelves containing FC disks and storage shelves containing ATA disks in the same loop. 12 | Physical Storage Management Guide
Methods of calculating aggregate and system capacity You use the physical and usable capacity of the drives you employ in your storage systems to ensure that your storage architecture conforms to the overall system capacity limits and the size limits of your aggregates. To maintain compatibility across different brands of drives, Data ONTAP rounds down (right-sizes) the amount of space available for user data. In addition, the numerical base used to calculate capacity (base 2 or base 10) also impacts sizing information. For these reasons, it is important to use the correct size measurement, depending on the task you want to accomplish: • For calculating overall system capacity, you use the physical capacity of the drive, and count every drive that is owned by the storage system. • For calculating how many drives you can put into an aggregate before you exceed its maximum size, you use the right-sized, or usable, capacity of all data drives in that aggregate. Parity, dparity, and cache drives are not counted against the maximum aggregate size. To see the physical and usable capacity for a specific drive, see the Hardware Universe at hwu.netapp.com. Disk speeds supported by Data ONTAP For hard disk drives, which use rotating media, speed is measured in revolutions per minute (RPM). Faster drives provide more input/output operations per second (IOPS) and faster response time. It is best to use disks of the same speed in an aggregate. Data ONTAP supports the following rotational speeds for hard disk drives: • Performance disks (SAS-connected) ◦ 10K RPM ◦ 15K RPM • Capacity disks (SAS-connected) ◦ 7.2K RPM • Performance disks (FC-connected) ◦ 15K RPM • Capacity disks (FC-connected) ◦ 7.2K RPM Managing disks using Data ONTAP | 13
Solid-state drives, or SSDs, are flash media-based devices and therefore the concept of rotational speed does not apply to them. For more information about which disks are supported with specific hardware configurations, see the Hardware Universe at hwu.netapp.com. How drive checksum types affect aggregate and spare management There are two checksum types available for drives used by Data ONTAP: BCS (block) and AZCS (zoned). Understanding how the checksum types differ and how they impact storage management enables you to manage your storage more effectively. Both checksum types provide the same resiliency capabilities. BCS optimizes for data access speed, and reserves the smallest amount of capacity for the checksum for drives with 520-byte sectors. AZCS provides enhanced storage utilization and capacity for drives with 512-byte sectors. You cannot change the checksum type of a drive. To determine the checksum type of a specific drive model, see the Hardware Universe. Aggregates have a checksum type, which is determined by the checksum type of the drives or array LUNs that compose the aggregate. The following configuration rules apply to aggregates, drives, and checksums: • Checksum types cannot be combined within RAID groups. This means that you must consider checksum type when you provide hot spare drives. • When you add storage to an aggregate, if it has a different checksum type than the storage in the RAID group to which it would normally be added, Data ONTAP creates a new RAID group. • An aggregate can have RAID groups of both checksum types. These aggregates have a checksum type of mixed. • For mirrored aggregates, both plexes must have the same checksum type. • Drives of a different checksum type cannot be used to replace a failed drive. • You cannot change the checksum type of a drive. Drive name formats When a node is part of a functioning cluster, the drives it is connected to are accessible using a simple, consistent drive name format. The drive name is independent of what nodes the drive is physically connected to and from which node you are accessing the drive. The drive name format is a concatenation of four components: <stack_id>.<shelf_id>.<bay>.<position> 14 | Physical Storage Management Guide
Note: During system boot, before the node has joined the cluster, or if certain key cluster components become unavailable, drive names revert to the classic format based on physical connectivity. • The stack ID is assigned by Data ONTAP. Stack IDs are unique across the cluster, and start with 1. • The shelf ID is set on the storage shelf when the shelf is added to the stack or loop. If there is a shelf ID conflict for SAS shelves, the shelf id is replaced with the shelf serial number in the drive name. • The bay is the position of the disk within its shelf. You can find the bay map in the administration guide for your storage shelf. • The position is used only for multi-disk carrier storage shelves. For carriers that house two disks, it can be 1 or 2. Related concepts Pre-cluster array LUN name format on page 74 Pre-cluster drive name formats on page 15 Pre-cluster drive name formats on page 15 Related references Commands for displaying information about storage shelves on page 52 Related information SAS Disk Shelves Installation and Service Guide for DS4243, DS2246, DS4486, and DS4246 DS14mk2 FC, and DS14mk4 FC Hardware Service Guide DiskShelf14mk2 AT Hardware Service Guide Pre-cluster drive name formats During steady-state operations, drive names are independent of their physical connection and unaffected by the node from which they are viewed. However, during system boot, before a node has joined a cluster, and if certain system components become unavailable, drive names revert to the format used before Data ONTAP 8.3, the pre-cluster format. The pre-cluster names of unowned drives (broken or unassigned drives) display the alphabetically lowest node name in the cluster that can see that drive. The following table shows the various formats for pre-cluster drive names, depending on how they are connected to the storage system. Managing disks using Data ONTAP | 15
Note: For internal drives, the slot number is zero, and the internal port number depends on the system model. Drive connection Drive name Example SAS, direct-attached <node>:<slot><port>.<shelfID>.<bay> The pre-cluster name for the drive in shelf 2, bay 11, connected to onboard port 0a and owned by node1 is node1:0a. 2.11. The pre-cluster name for the drive in shelf 6, bay 3, connected to an HBA in slot 1, port c, and owned by node1 is node1:1c. 6.3. SAS, direct-attached in multi-disk carrier disk shelf <node>:<slot><port>.<shelfID>.<bay>L<carrierP osition> Carrier position is 1 or 2. SAS, direct-attached, for systems running Data ONTAP-v <slot><port>.<ID> The pre-cluster name for the third virtual disk connected to the first port is 0b.3. The pre-cluster name for the second virtual disk connected to the third port is 0d.2. The range of ports is b through e, and the range of disks is 0 through 15. 16 | Physical Storage Management Guide
Drive connection Drive name Example SAS, bridge-attached (FibreBridge, used for MetroCluster configurations) <slot><port>.<loopID>L<LUN> The pre-cluster name for the drive with LUN 2 behind the bridge connected to port a in slot 3, loop ID 125, is 3a.125L2. SAS, bridge-attached through a switch (FibreBridge, used for MetroCluster configurations) <switch_name>:<switch_port>.<loopID>L<LUN > The pre-cluster name for the drive with LUN 5 behind the bridge connected to port 2 of switch brcd44, loop ID 126, is brcd44:2.126L5. FC, direct-attached <node>:<slot><port>.<loopID> The pre-cluster name for the drive with loop ID 19 (bay 3 of shelf 1) connected to onboard port 0a and owned by node1 is node1:0a. 19. The pre-cluster name for the drive with loop ID 34 connected to an HBA in slot 8, port c and owned by node1 is node1:8c. 34. FC, switch-attached <node>:<switch_name>.<switch_port>.<loopID> The pre-cluster name for the drive with loop ID 51 connected to port 3 of switch SW7 owned by node1 is node1:SW7.3.51. Managing disks using Data ONTAP | 17
Each drive has a universal unique identifier (UUID) that differentiates it from all other drives in the cluster. Related concepts Drive name formats on page 14 Pre-cluster array LUN name format on page 74 Loop IDs for FC-AL connected disks For disks connected using Fibre Channel-Arbitrated Loop (FC-AL or FC), the loop ID is an integer between 16 and 126. The loop ID identifies the disk within its loop, and is included in the disk name, which identifies the disk uniquely for the entire system. The loop ID corresponds to the storage shelf number and the bay in which the disk is installed. The lowest loop ID is always in the far right bay of the first storage shelf. The next higher loop ID is in the next bay to the left, and so on. You can view the device map for your storage shelves by using the fcadmin device_map command, available through the nodeshell. For more information about the loop ID map for your storage shelf, see the hardware guide for the storage shelf. Understanding RAID drive types Data ONTAP classifies drives (or, for partitioned drives, partitions) as one of four types for RAID: data, hot spare, parity, or dParity. You manage disks differently depending on whether they are spare or being used in an aggregate. The RAID type is determined by how RAID is using a drive or partition; it is different from the Data ONTAP disk type. You cannot affect the RAID type for a drive. The RAID type is displayed in the Position column for many storage commands. For drives using root-data partitioning and SSDs in storage pools, a single drive might be used in multiple ways for RAID. For example, the root partition of a partitioned drive might be a spare partition, whereas the data partition might be being used for parity. For this reason, the RAID drive type for partitioned drives and SSDs in storage pools is displayed simply as shared. Data disk Holds data stored on behalf of clients within RAID groups (and any data generated about the state of the storage system as a result of a malfunction). Spare disk Does not hold usable data, but is available to be added to a RAID group in an aggregate. Any functioning disk that is not assigned to an aggregate but is assigned to a system functions as a hot spare disk. 18 | Physical Storage Management Guide
Parity disk Stores row parity information that is used for data reconstruction when a single disk drive fails within the RAID group. dParity disk Stores diagonal parity information that is used for data reconstruction when two disk drives fail within the RAID group, if RAID-DP is enabled. Related concepts How Data ONTAP reports disk types on page 10 How disk sanitization works Disk sanitization is the process of physically obliterating data by overwriting disks or SSDs with specified byte patterns or random data so that recovery of the original data becomes impossible. You use the sanitization process to ensure that no one can recover the data on the disks. This functionality is available through the nodeshell. Related tasks Using disk sanitization to remove data from disks on page 48 Disk sanitization process Understanding the basics of the disk sanitization process helps you understand what to anticipate during the sanitization process and after it is complete. The disk sanitization process uses three successive default or user-specified byte overwrite patterns for up to seven cycles per operation. The random overwrite pattern is repeated for each cycle. Depending on the disk capacity, the patterns, and the number of cycles, the process can take several hours. Sanitization runs in the background. You can start, stop, and display the status of the sanitization process. The sanitization process contains two phases: 1. Formatting phase The operation performed for the formatting phase depends on the class of disk being sanitized, as shown in the following table: Disk class Formatting phase Capacity HDDs Skipped Performance HDDs SCSI format operation SSDs SCSI sanitize operation Managing disks using Data ONTAP | 19
2. Pattern overwrite phase The specified overwrite patterns are repeated for the specified number of cycles. When the sanitization process is complete, the specified disks are in a sanitized state. They are not returned to spare status automatically. You must return the sanitized disks to the spare pool before the newly sanitized disks are available to be added to another aggregate. When disk sanitization cannot be performed Disk sanitization is not supported for all disk types. In addition, there are times when disk sanitization cannot be performed. You should be aware of the following facts about the disk sanitization process: • It is not supported on all SSD part numbers. For information about which SSD part numbers support disk sanitization, see the Hardware Universe at hwu.netapp.com. • It is not supported in takeover mode for systems in an HA pair. • It cannot be performed on disks that were failed due to readability or writability problems. • It does not perform its formatting phase on ATA drives. • If you are using the random pattern, it cannot be performed on more than 100 disks at one time. • It is not supported on array LUNs. • If you sanitize both SES disks in the same ESH shelf at the same time, you see errors on the console about access to that shelf, and shelf warnings are not reported for the duration of the sanitization. However, data access to that shelf is not interrupted. • You can perform disk sanitization on disks using Storage Encryption. However, there are other methods to obliterate data on disks using Storage Encryption that are faster and do not require an operational storage system. What happens if disk sanitization is interrupted Disk sanitization is a long-running operation. If disk sanitization is interrupted by user intervention or an unexpected event such as a power outage, Data ONTAP takes action to return the disks that were being sanitized to a known state, but you must also take action before the sanitization process can finish. If the sanitization process is interrupted by power failure, system panic, or manual intervention, the sanitization process must be repeated from the beginning. The disk is not designated as sanitized. If the formatting phase of disk sanitization is interrupted, Data ONTAP must recover any disks that were corrupted by the interruption. After a system reboot and once every hour, Data ONTAP checks for any sanitization target disk that did not complete the formatting phase of its sanitization. If any 20 | Physical Storage Management Guide
such disks are found, Data ONTAP recovers them. The recovery method depends on the type of the disk. After a disk is recovered, you can rerun the sanitization process on that disk; for HDDs, you can use the -s option to specify that the formatting phase is not repeated again. Tips for creating and backing up aggregates containing data to be sanitized If you are creating or backing up aggregates to contain data that might need to be sanitized, following some simple guidelines will reduce the time it takes to sanitize your data. • Make sure your aggregates containing sensitive data are not larger than they need to be. If they are larger than needed, sanitization requires more time, disk space, and bandwidth. • When you back up aggregates containing sensitive data, avoid backing them up to aggregates that also contain large amounts of nonsensitive data. This reduces the resources required to move nonsensitive data before sanitizing sensitive data. How Data ONTAP monitors disk performance and health Data ONTAP continually monitors disks to assess their performance and health. When Data ONTAP encounters certain errors or behaviors from a disk, it takes the disk offline temporarily or takes the disk out of service to run further tests. What happens when Data ONTAP takes disks offline Data ONTAP temporarily stops I/O activity to a disk and takes a disk offline when Data ONTAP is updating disk firmware in background mode or when disks become non-responsive. While the disk is offline, Data ONTAP performs a quick check on it to reduce the likelihood of forced disk failures. A disk can be taken offline only if its containing RAID group is in a normal state and the plex or aggregate is not offline. While the disk is offline, Data ONTAP reads from other disks within the RAID group while writes are logged. When the offline disk is ready to come back online, Data ONTAP resynchronizes the RAID group and brings the disk online. This process generally takes a few minutes and incurs a negligible performance impact. How Data ONTAP reduces disk failures using Rapid RAID Recovery When Data ONTAP determines that a disk has exceeded its error thresholds, Data ONTAP can perform Rapid RAID Recovery by removing the disk from its RAID group for testing and, if necessary, failing the disk. Spotting disk errors quickly helps prevent multiple disk failures and allows problem disks to be replaced. By performing the Rapid RAID Recovery process on a suspect disk, Data ONTAP avoids three problems that occur during sudden disk failure and the subsequent RAID reconstruction process: • Rebuild time Managing disks using Data ONTAP | 21
• Performance degradation • Potential data loss due to additional disk failure during reconstruction During Rapid RAID Recovery, Data ONTAP performs the following tasks: 1. Places the suspect disk in pre-fail mode. 2. Selects a hot spare replacement disk. Note: If no appropriate hot spare is available, the suspect disk remains in pre-fail mode and data continues to be served. However, a suspect disk performs less efficiently. Impact on performance ranges from negligible to worse than degraded mode. For this reason, hot spares should always be available. 3. Copies the suspect disk’s contents to the spare disk on the storage system before an actual failure occurs. 4. After the copy is complete, attempts to put the suspect disk into the maintenance center, or else fails the disk. Note: Tasks 2 through 4 can occur only when the RAID group is in normal (not degraded) mode. If the suspect disk fails on its own before copying to a hot spare is complete, Data ONTAP starts the normal RAID reconstruction process. A message is sent to the log file when the Rapid RAID Recovery process is started and when it is complete. The messages are tagged raid.rg.diskcopy.start:notice and raid.rg.diskcopy.done:notice. Related concepts About degraded mode on page 119 When Data ONTAP can put a disk into the maintenance center on page 23 How Data ONTAP works with hot spare disks on page 117 How the maintenance center helps prevent drive errors Drives can sometimes display small problems that do not interfere with normal operation, but which could be a sign that the drive might fail sometime soon. The maintenance center provides a way to put these drives under increased scrutiny. When a suspect drive is in the maintenance center, it is subjected to a number of tests. If the drive passes all of the tests, Data ONTAP redesignates it as a spare; if it fails any tests, Data ONTAP fails the drive. By default, Data ONTAP puts a suspect drive into the maintenance center automatically only if there are two or more spares available for that drive. If that drive is housed in a shelf that supports automatic power-cycling, power to that drive might be turned off for a short time. If the drive returns 22 | Physical Storage Management Guide
to a ready state after the power-cycle, the maintenance center tests the drive. Otherwise, the maintenance center fails the drive immediately. You can put a suspect drive into the maintenance center manually regardless of how many spares are available. You can also specify the number of times a drive is allowed to go to the maintenance center and whether it should go immediately or only after copying its contents to a spare drive. • The disk.maint_center.enable option controls whether the maintenance center is on or off. The default value is on. • The disk.maint_center.allowed_entries option controls the number of times a suspect drive is allowed to go to the maintenance center. The default value is 1, which means that if the drive is sent to the maintenance center more than once, it is automatically failed. • The disk.maint_center.spares_check option controls whether Data ONTAP requires that sufficient spares are available before it puts a drive into the maintenance center. • The disk maint start command, available through the nodeshell, enables you to put a suspect drive into the maintenance center manually. If the target drive is in use, the default is that it does not enter the maintenance center until its contents have been copied to a spare drive. You can put the drive into the maintenance center immediately by including the -i option. Related information NetApp Technical Report 3437: Storage Subsystem Resiliency Guide When Data ONTAP can put a disk into the maintenance center When Data ONTAP detects certain disk errors, it tries to put the disk into the maintenance center for testing. Certain requirements must be met for the disk to be put into the maintenance center. If a disk experiences more errors than are allowed for that disk type, Data ONTAP takes one of the following actions: • If the disk.maint_center.spares_check option is set to on (the default) and two or more spares are available (four for multi-disk carriers), Data ONTAP takes the disk out of service and assigns it to the maintenance center for data management operations and further testing. • If the disk.maint_center.spares_check option is set to on and fewer than two spares are available (four for multi-disk carriers), Data ONTAP does not assign the disk to the maintenance center. It fails the disk and designates the disk as a broken disk. • If the disk.maint_center.spares_check option is set to off, Data ONTAP assigns the disk to the maintenance center without checking the number of available spares. Managing disks using Data ONTAP | 23
Note: The disk.maint_center.spares_check option has no effect on putting disks into the maintenance center from the command-line interface. Data ONTAP does not put SSDs into the maintenance center. How Data ONTAP uses continuous media scrubbing to prevent media errors The purpose of the continuous media scrub is to detect and correct media errors to minimize the chance of storage system disruption due to a media error while a storage system is in degraded or reconstruction mode. By default, Data ONTAP runs continuous background media scrubbing for media errors on all storage system disks. If a media error is found, Data ONTAP uses RAID to reconstruct the data and repairs the error. Media scrubbing is a continuous background process. Therefore, you might observe disk LEDs blinking on an apparently idle storage system. You might also observe some CPU activity even when no user workload is present. How continuous media scrubbing impacts system performance Because continuous media scrubbing searches only for media errors, its impact on system performance is negligible. In addition, the media scrub attempts to exploit idle disk bandwidth and free CPU cycles to make faster progress. However, any client workload results in aggressive throttling of the media scrub resource. If needed, you can further decrease the CPU resources consumed by a continuous media scrub under a heavy client workload by increasing the maximum time allowed for a media scrub cycle to complete. You can do this by using the raid.media_scrub.rate option. Why continuous media scrubbing should not replace scheduled RAID-level disk scrubs Because the continuous media scrub process scrubs only media errors, you should continue to run the storage system’s scheduled complete RAID-level scrub operation. The RAID-level scrub finds and corrects parity and checksum errors as well as media errors. How you can use ACP to increase storage availability for SAS-connected disk shelves The Alternate Control Path (ACP) protocol enables Data ONTAP to manage and control SAS- connected storage shelves, which can increase storage availability. After you ensure that ACP is properly enabled and configured, you do not need to actively manage it. You can install SAS-connected storage shelves without configuring ACP. However, for maximum storage availability and stability, you should always have ACP enabled and configured, including 24 | Physical Storage Management Guide
providing the required physical connectivity and configuration parameters to enable the ACP functionality. ACP is enabled by default. If you need to enable it or change its configuration, you can use the acpadmin configure and storage show acp commands, available through the nodeshell. You do not need to actively manage the SAS-connected storage shelf subsystem. Data ONTAP automatically monitors and manages the subsystem without operator intervention. How you use SSDs to increase storage performance Solid-state drives (SSDs) are flash media-based storage devices that provide better overall performance than hard disk drives (HDDs), which are mechanical devices using rotating media. You should understand how Data ONTAP manages SSDs and the capability differences between SSDs and HDDs. Depending on your storage system model, you can use SSDs in two ways: • You can create Flash Pool aggregates—aggregates composed mostly of HDDs, but with some SSDs that function as a high-performance cache for your working data set. • You can create aggregates composed entirely of SSDs, where the SSDs function as the persistent storage for all data in the aggregate. You manage Flash Pool aggregates and aggregates composed entirely of SSDs the same way you manage aggregates composed entirely of HDDs. However, there are some differences in the way you manage SSDs from the way you manage disks. In addition, some Data ONTAP capabilities are not available on SSDs and Flash Pool aggregates. Related concepts How Flash Pool aggregates work on page 130 How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates on page 137 How the All-Flash Optimized personality affects node behavior When you order an All-Flash FAS platform model, it has the All-Flash Optimized personality. Having the All-Flash Optimized personality enabled for a node introduces some extra requirements for that node. When the All-Flash Optimized personality is enabled, you cannot attach HDDs or array LUNs to the node; no drives other than SSDs are recognized as supported. This enables the node to provide performance that can be achieved only with a pure SSD solution. Both partners in an HA pair must have the same All-Flash Optimized personality (either enabled or disabled). You cannot move an aggregate that contains HDD RAID groups to a node with the All- Managing disks using Data ONTAP | 25
Flash Optimized personality. Volume move is unaffected; you can move volumes between nodes with and nodes without the All-Flash Optimized personality. Root-data partitioning is enabled by default on any node with the All-Flash Optimized personality. You can determine whether a node has the All-Flash Optimized personality by using the system node show command and looking for All-Flash Optimized. How Data ONTAP manages SSD wear life Solid-state disks (SSDs) have a different end-of-life behavior than rotating media (hard disk drives, or HDDs). Data ONTAP monitors and manages SSDs to maximize storage performance and availability. In the absence of a mechanical failure, rotating media can serve data almost indefinitely. This is not true for SSDs, which can accept only a finite (though very large) number of write operations. SSDs provide a set of internal spare capacity, called spare blocks, that can be used to replace blocks that have reached their write operation limit. After all of the spare blocks have been used, the next block that reaches its limit causes the disk to fail. Because a drive failure is an undesirable occurrence, Data ONTAP replaces SSDs before they reach their limit. When a predetermined percentage of the spare blocks have been used (approximately 90%), Data ONTAP performs the following actions: 1. Sends an AutoSupport message. 2. If a spare SSD is available, starts a disk copy to that spare. 3. If no spare is available, starts a periodic check for a spare so that the disk copy can be started when a spare becomes available. 4. When the disk copy finishes, fails the disk. Note: You do not need to replace SSDs before they are failed by Data ONTAP. However, when you use SSDs in your storage system (as for all disk types), it is important to ensure that you have sufficient hot spares available at all times. Capability differences between SSDs and HDDs Usually, you manage SSDs the same as HDDs, including firmware updates, scrubs, and zeroing. However, some Data ONTAP capabilities do not make sense for SSDs, and SSDs are not supported on all hardware models. SSDs cannot be combined with HDDs within the same RAID group. When you replace an SSD in an aggregate, you must replace it with another SSD. Similarly, when you physically replace an SSD within a shelf, you must replace it with another SSD. The following capabilities of Data ONTAP are not available for SSDs: • Disk sanitization is not supported for all SSD part numbers. 26 | Physical Storage Management Guide
• The maintenance center Guidelines and requirements for using multi-disk carrier storage shelves Data ONTAP automatically handles most of the extra steps required to manage disks in multi-disk carriers. However, there are some extra management and configuration requirements that you must understand before incorporating multi-disk carrier disk shelves in your storage architecture. When using storage from multi-disk carrier disk shelves such as the DS4486, you must familiarize yourself with the guidelines and requirements governing the following topics: • The process that Data ONTAP uses to avoid impacting any RAID groups when a multi-disk carrier needs to be removed • When it is safe to remove a multi-disk carrier after a disk failure • The minimum required number of spares for multi-disk carrier disks • Multi-disk carrier disk shelf configuration • Aggregate configuration requirements when using multi-disk carrier disk shelves • Guidelines and best practices for using disks from a multi-disk carrier disk shelf in an aggregate How Data ONTAP avoids RAID impact when a multi-disk carrier must be removed Data ONTAP takes extra steps to ensure that both disks in a carrier can be replaced without impacting any RAID group. Understanding this process helps you know what to expect when a disk from a multi-disk carrier storage shelf fails. A multi-disk carrier storage shelf, such as the DS4486, has double the storage density of other SAS- connected storage shelves. It accomplishes this by housing two disks per disk carrier. When two disks share the same disk carrier, they must be removed and inserted together. This means that when one of the disks in a carrier needs to be replaced, the other disk in the carrier must also be replaced, even if it was not experiencing any issues. Removing two data or parity disks from an aggregate at the same time is undesirable, because it could leave two RAID groups degraded, or one RAID group double-degraded. To avoid this situation, Data ONTAP initiates a storage evacuation operation for the carrier mate of the failed disk, as well as the usual reconstruction to replace the failed disk. The disk evacuation operation copies the contents of the carrier mate to a disk in a different carrier so that the data on that disk remains available when you remove the carrier. During the evacuation operation, the status for the disk being evacuated is shown as evacuating. In addition, Data ONTAP tries to create an optimal layout that avoids having two carrier mates in the same RAID group. Depending on how the other disks are laid out, achieving the optimal layout can Managing disks using Data ONTAP | 27
require as many as three consecutive disk evacuation operations. Depending on the size of the disks and the storage system load, each storage evacuation operation could take several hours, so the entire swapping process could take an entire day or more. If insufficient spares are available to support the swapping operation, Data ONTAP issues a warning and waits to perform the swap until you provide enough spares. How to determine when it is safe to remove a multi-disk carrier Removing a multi-disk carrier before it is safe to do so can result in one or more RAID groups becoming degraded, or possibly even a storage disruption. Data ONTAP provides several indications of when it is safe to remove a multi-disk carrier. When a multi-disk carrier needs to be replaced, the following events must have occurred before you can remove the carrier safely: • An AutoSupport message must have been logged indicating that the carrier is ready to be removed. • An EMS message must have been logged indicating that the carrier is ready to be removed. • Both disks in the carrier must be displayed in the list of broken disks. You can see the list of broken disks by using the storage disk show -broken command. The disk that was evacuated to allow the carrier to be removed shows the outage reason of evacuated. • The amber LED on the carrier must be lit continuously. • The green LED on the carrier must show no activity. Attention: You cannot reuse the carrier mate of a failed disk. When you remove a multi-disk carrier that contains a failed disk, you must replace and return the entire carrier. Spare requirements for multi-disk carrier disks Maintaining the proper number of spares for disks in multi-disk carriers is critical for optimizing storage redundancy and minimizing the amount of time Data ONTAP must spend copying disks to achieve an optimal disk layout. You must maintain a minimum of two hot spares for multi-disk carrier disks at all times. To support the use of the Maintenance Center, and to avoid issues caused by multiple concurrent disk failures, you should maintain at least four hot spares for steady state operation, and replace failed disks promptly. If two disks fail at the same time with only two available hot spares, Data ONTAP might not be able to swap the contents of both the failed disk and its carrier mate to the spare disks. This scenario is called a stalemate. If this happens, you are notified through EMS messages and AutoSupport messages. When the replacement carriers become available, you must follow the instructions provided by the EMS messages or contact technical support to recover from the stalemate. 28 | Physical Storage Management Guide
Shelf configuration requirements for multi-disk carrier storage shelves You can combine multi-disk carrier disk shelves with single-disk carrier disk shelves (standard disk shelves) on the same storage system and within in the same stack. Aggregate requirements for disks in multi-disk carrier storage shelves Aggregates composed of disks in multi-disk carrier disk shelves must conform to some configuration requirements. The following configuration requirements apply to aggregates composed of disks in multi-disk carrier disk shelves: • The RAID type must be RAID-DP. • All HDDs in the aggregate must be the same Data ONTAP disk type. The aggregate can be a Flash Pool aggregate. • If the aggregate is mirrored, both plexes must have the same Data ONTAP disk type (or types, in the case of a Flash Pool aggregate). Related concepts How Flash Pool aggregates work on page 130 Considerations for using disks from a multi-disk carrier storage shelf in an aggregate Observing the requirements and best practices for using disks from a multi-disk carrier disk shelf in an aggregate enables you to maximize storage redundancy and minimize the impact of disk failures. Disks in multi-disk carriers always have the Data ONTAP disk type of MSATA. MSATA disks cannot be mixed with HDDs from a single-carrier disk shelf in the same aggregate. The following disk layout requirements apply when you are creating or increasing the size of an aggregate composed of MSATA disks: • Data ONTAP prevents you from putting two disks in the same carrier into the same RAID group. • Do not put two disks in the same carrier into different pools, even if the shelf is supplying disks to both pools. • Do not assign disks in the same carrier to different nodes. • For the best layout, do not name specific disks; allow Data ONTAP to select the disks to be used or added. If the operation cannot result in an optimal layout due to the placement of the disks and available spares, Data ONTAP automatically swaps disk contents until an optimal layout is achieved. If there are not enough available spares to support the swaps, Data ONTAP issues a warning and Managing disks using Data ONTAP | 29
waits to perform the disk swaps until you provide the necessary number of hot spares. If you name disks and an optimal layout cannot be achieved, you must explicitly force the operation; otherwise, the operation fails. Aggregate creation example To create an aggregate using MSATA disks, you can specify the disk type and size but leave the disk selection and layout to Data ONTAP by using a command like this: storage aggregate create -aggregate c1n1_aggr1 -node node1 - disktype MSATA -diskcount 14 Understanding root-data partitioning Some platform models use partioning to enable the root aggregate to use less space, leaving more space for the data aggregate and improving storage utilization. Root-data partitioning, also called shared drives, changes the way you view and administer your disks and aggregates. How root-data partitioning works For entry-level and All Flash FAS (AFF) platform models, aggregates can be composed of parts of a drive rather than the entire drive. Root-data partitioning is usually enabled and configured by the factory. It can also be established by initiating system initialization using option 4 from the boot menu. Note that system initialization erases all data on the disks of the node and resets the node configuration to the factory default settings. When a node has been configured to use root-data partitioning, partitioned disks have two partitions: The smaller partition is used to compose the root aggregate. The larger partition is used in data aggregates. The size of the partitions is set by Data ONTAP, and depends on the number of disks used to compose the root aggregate when the system is initialized. (The more disks used to compose the root aggregate, the smaller the root partition.) After system initialization, the partition sizes are fixed; adding partitions or disks to the root aggregate after system initialization increases the size of the root aggregate, but does not change the root partition size. 30 | Physical Storage Management Guide
The partitions are used by RAID in the same manner as physical disks are; all of the same requirements apply. For example, if you add an unpartitioned drive to a RAID group consisting of partitioned drives, the unpartitioned drive is partitioned to match the partition size of the drives in the RAID group and the rest of the disk is unused. If a partitioned disk is moved to another node or used in another aggregate, the partitioning persists; you can use the disk only in RAID groups composed of partitioned disks. Related information Clustered Data ONTAP 8.3 System Administration Guide for Cluster Administrators How root-data partitioning affects storage management Generally, you administer a node that is using root-data partitioning the same way you administer a node that uses only physical disks. However, partitioned disks are displayed slightly differently, and you need to take their presence into account for some management tasks. Root and data partitions are used by the RAID subsystem the same way that physical disks are, and the same requirements and best practices apply. You follow the same best practices for spare management, and disk names are not impacted by root-data partitioning. When you are working with a node that is using root-data partitioning, the following tasks are slightly different: • Creating and adding storage to aggregates Spare partitions are displayed differently—as either root or data. In addition, you need to ensure that a root and data spare are provided on the same disk to support core file creation. Finally, you should understand whether you are mixing physical disks and partitions in the same RAID group, and the impact of doing so. • Displaying drive and RAID group information for an aggregate Partitioned disks display shared for the Position column, rather than their RAID disk type. • Assigning and removing ownership Automatic ownership assignment works for partitioned disks. If you need to manually assign or remove ownership for partitioned disks, you must do so for both partitions as well as the container disk. • Removing and replacing disks Removing a partitioned disk can potentially affect more than one aggregate. Related concepts Which drives are partitioned and used for the root aggregate on page 32 Removing disks from a node on page 45 Requirements for using root-data partitioning on page 34 Managing disks using Data ONTAP | 31
Related tasks Creating an aggregate using root-data partitioning on page 161 Increasing the size of an aggregate that uses root-data partitioning on page 165 Determining drive and RAID group information for an aggregate on page 181 Assigning ownership for disks partitioned for root-data partitioning on page 59 Related references Commands for managing disks on page 51 Which drives are partitioned and used for the root aggregate The drives that are partitioned for use in the root aggregate depend on the system configuration. Knowing how many drives are used for the root aggregate helps you to determine how much of the drives' capacity is reserved for the root partition, and how much is available for use in a data aggregate. Root-data partitioning is supported for two types of platform: entry-level platforms, and All-Flash FAS (AFF) platforms. For entry-level platforms, only the internal drives are partitioned. For AFF platforms, all drives that are attached to the controller when the system is initialized are partitioned, up to a limit of 24 per node. Drives that are added after system configuration are not partitioned. Standard root-data partitioning layouts The root aggregate is configured by the factory; you should not change it. However, you can use the data partitions in a few different configurations, depending on your requirements. The following diagram shows one way to configure the partitions for an active-passive configuration with 12 partitioned disks. There are two root aggregates, one for each node, composed of the small partitions. Each root aggregate has a spare partition. There is just one RAID-DP data aggregate, with two parity disk partitions and one spare partition. 32 | Physical Storage Management Guide
The following diagram shows one way to configure the partitions for an active-active configuration with 12 partitioned disks. In this case, there are two RAID-DP data aggregates, each with their own data partitions, parity partitions, and spares. Note that each disk is allocated to only one node. This is a best practice that prevents the loss of a single disk from affecting both nodes. The disks used for data, parity, and spare partitions might not be exactly as shown in these diagrams. For example, the parity partitions might not always align on the same disk. Managing disks using Data ONTAP | 33
Requirements for using root-data partitioning In most cases, you can use drives that are partitioned for root-data partitioning exactly as you would use a physical, unshared drive. However, you cannot use root-data partitioning in certain configurations. The following storage devices cannot be partitioned: • Array LUNs • Virtual disks as created by Data ONTAP-v • HDD types that are not available as internal drives: ATA, FCAL, and MSATA • 100-GB SSDs You cannot use root-data partitioning with the following technologies: • MetroCluster • Data ONTAP-v • RAID4 Aggregates composed of partitioned drives must have a RAID type of RAID-DP. Initializing a node to configure root-data partitioning If you are initializing a node whose platform model supports root-data partitioning, you must complete some steps before performing the initialization to ensure that root-data partitioning is configured correctly. Before you begin • Attention: All data that you need to retain must have been migrated to another node. All data on the HA pair is erased during this procedure and cannot be retrieved. • Your node configuration must support root-data partitioning. If the node does not satisfy the requirements for root-data partitioning, root-data partitioning is not configured. • The node or HA pair you are initializing should not be joined with more nodes in a cluster, and should not be serving data. • Your nodes must be running Data ONTAP 8.3 or later. • Neither node is in takeover or giveback. 34 | Physical Storage Management Guide
About this task This procedure can be used for both entry-level platform models and All-Flash FAS (AFF) platform models. For entry-level platform models, only internal disks are partitioned. For AFF platform models, up to 48 SSDs are partitioned, depending on the number of SSDs that are attached to the controller. This procedure is designed to be run on an HA pair, so the nodes are called Node A and Node B. However, you can still use this procedure if you have only one controller; ignore instructions that are specifically for Node B. This procedure can take many hours to complete, depending on the amount and type of storage attached to the HA pair. Steps 1. Record your node configuration, including network configuration, license values, and passwords, and if you want to restore the same storage architecture as you have now, record that also. All of the current node configuration information will be erased when the system is initialized. 2. If you are performing this procedure on a two-node cluster, disable the HA capability: cluster ha modify -configured false This can be done from either node. 3. Disable storage failover for both nodes: storage failover modify -enabled false -node nodeA,nodeB 4. On both nodes, boot into maintenance mode: a. Halt the system: system node halt -node node_name You can ignore any error messages about epsilon. For the second node, you can include the - skip-lif-migration-before-shutdown flag if prompted to do so. b. At the LOADER prompt, boot Data ONTAP: boot_ontap c. Monitor the boot process, and when prompted, press Ctrl-C to display the boot menu. Managing disks using Data ONTAP | 35
Example ******************************* * * * Press Ctrl-C for Boot Menu. * * * ******************************* d. Select option 5, Maintenance mode boot. Example Please choose one of the following: (1) Normal Boot. (2) Boot without /etc/rc. (3) Change password. (4) Clean configuration and initialize all disks. (5) Maintenance mode boot. (6) Update flash from backup config. (7) Install new software first. (8) Reboot node. Selection (1-8)? 5 5. For both nodes, if there are external disks connected to the node, destroy all aggregates, including the root aggregate: a. Display all aggregates: aggr status b. For each aggregate, take the aggregate offline: aggr offline aggr_name c. For each aggregate, destroy the aggregate: aggr destroy aggr_name All disks are converted to spares. 6. Ensure that no drives in the HA pair are partitioned: a. Display all drives owned by both nodes: disk show If any of the drives show three entries, that drive has been partitioned. If your nodes have no partitioned disks, you can skip to step 7. The following example shows partitioned drive 0a.10.11: 36 | Physical Storage Management Guide
Example DISK OWNER POOL SERIAL NUMBER HOME ------------ ------------- ----- ------------- ------------- 0a.10.11 sys1(536880559) Pool0 N11YE08L sys1(536880559) 0a.10.11P1 sys1(536880559) Pool0 N11YE08LNP001 sys1(536880559) 0a.10.11P2 sys1(536880559) Pool0 N11YE08LNP002 sys1(536880559) b. Note any partitions that do not have the same owner as their container disk. Example In the following example, the container disk (0a.10.14) is owned by sys1, but partition one (0a.10.14P1) is owned by sys2. DISK OWNER POOL SERIAL NUMBER HOME ------------ ------------- ----- ------------- ------------- 0a.10.14 sys1(536880559) Pool0 N11YE08L sys1(536880559) 0a.10.14P1 sys2(536880408) Pool0 N11YE08LNP001 sys2(536880408) 0a.10.14P2 sys1(536880559) Pool0 N11YE08LNP002 sys1(536880559) c. Update the ownership of all partitions owned by a different node than their container disk: disk assign disk_partition -f -o container_disk_owner Example You would enter a command like the following example for each disk with partitions owned by a different node than their container disk: disk assign 0a.10.14P1 -f -o sys1 d. For each drive with partitions, on the node that owns the container disk, remove the partitions: disk unpartition disk_name disk_name is the name of the disk, without any partition information, such as “0a.10.3”. Example You would enter a command like the following example for each disk with partitions: disk unpartition 0a.10.14 7. On both nodes, remove disk ownership: disk remove_ownership 8. On both nodes, verify that all drives connected to both nodes are unowned: disk show Managing disks using Data ONTAP | 37
Example *> disk show Local System ID: 465245905 disk show: No disk match option show. 9. On both nodes, return to the LOADER menu: halt 10. On Node A only, begin zeroing the drives: a. At the LOADER prompt, boot Data ONTAP: boot_ontap b. Monitor the boot process, and when prompted, press Ctrl-C to display the boot menu. c. Select option 4, Clean configuration and initialize all disks. When the drives begin to be zeroed, a series of dots are printed to the console. 11. After the drives on Node A have been zeroing for a few minutes, repeat the previous step on Node B. The drives on both nodes that will be partitioned are zeroing . When the zeroing process is complete, the nodes return to the Node Setup wizard. 12. Restore your system configuration. 13. Confirm that the root aggregate on both nodes is composed of partitions: storage aggregate show-status The Position column shows shared and the usable size is a small fraction of the physical size: Owner Node: sys1 Aggregate: aggr0_1 (online, raid_dp) (block checksums) Plex: /aggr0_1/plex0 (online, normal, active, pool0) RAID Group /aggr0_1/plex0/rg0 (normal, block checksums) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- -------- shared 0a.10.10 0 BSAS 7200 73.89GB 828.0GB (normal) shared 0a.10.0 0 BSAS 7200 73.89GB 828.0GB (normal) shared 0a.10.11 0 BSAS 7200 73.89GB 828.0GB (normal) shared 0a.10.6 0 BSAS 7200 73.89GB 828.0GB (normal) shared 0a.10.5 0 BSAS 7200 73.89GB 828.0GB (normal) 14. Run System Setup to reconfigure the HA pair or rejoin the cluster, depending on your initial configuration. Related concepts Which drives are partitioned and used for the root aggregate on page 32 Requirements for using root-data partitioning on page 34 38 | Physical Storage Management Guide
Related information Clustered Data ONTAP 8.3 Software Setup Guide Clustered Data ONTAP 8.3 High-Availability Configuration Guide Setting up an active-passive configuration on nodes using root-data partitioning When an HA pair is configured to use root-data partitioning by the factory, ownership of the data partitions is split between both nodes in the pair, for use in an active-active configuration. If you want to use the HA pair in an active-passive configuration, you must update partition ownership before creating your data aggregate. Before you begin • You should have decided which node will be the active node and which node will be the passive node. • Storage failover must be configured on the HA pair. About this task This task is performed on two nodes: Node A and Node B. All commands are input at the clustershell. This procedure is designed for nodes for which no data aggregate has been created from the partitioned disks. Steps 1. View the current ownership of the data partitions: storage aggregate show-spare-disks Example You can see that half of the data partitions are owned by one node and half are owned by the other node. All of the data partitions should be spare. cluster1::> storage aggregate show-spare-disks Original Owner: cluster1-01 Pool0 Partitioned Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size --------------------------- ----- ------ -------------- -------- -------- -------- 1.0.0 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.1 BSAS 7200 block 753.8GB 73.89GB 828.0GB 1.0.5 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.6 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.10 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.11 BSAS 7200 block 753.8GB 0B 828.0GB Managing disks using Data ONTAP | 39
Original Owner: cluster1-02 Pool0 Partitioned Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size --------------------------- ----- ------ -------------- -------- -------- -------- 1.0.2 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.3 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.4 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.7 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.8 BSAS 7200 block 753.8GB 73.89GB 828.0GB 1.0.9 BSAS 7200 block 753.8GB 0B 828.0GB 12 entries were displayed. 2. Enter the advanced privilege level: set advanced 3. For each data partition owned by the node that will be the passive node, assign it to the active node: storage disk assign -force -data true -owner active_node_name -disk disk_name You do not need to include the partition as part of the disk name. Example You would enter a command similar to the following example for each data partition you need to reassign: storage disk assign -force -data true -owner cluster1-01 -disk 1.0.3 4. Confirm that all of the partitions are assigned to the active node. Example cluster1::*> storage aggregate show-spare-disks Original Owner: cluster1-01 Pool0 Partitioned Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size --------------------------- ----- ------ -------------- -------- -------- -------- 1.0.0 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.1 BSAS 7200 block 753.8GB 73.89GB 828.0GB 1.0.2 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.3 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.4 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.5 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.6 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.7 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.8 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.9 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.10 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.11 BSAS 7200 block 753.8GB 0B 828.0GB Original Owner: cluster1-02 Pool0 Partitioned Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size 40 | Physical Storage Management Guide
--------------------------- ----- ------ -------------- -------- -------- -------- 1.0.8 BSAS 7200 block 0B 73.89GB 828.0GB 13 entries were displayed. Note that cluster1-02 still owns a spare root partition. 5. Return to administrative privilege: set admin 6. Create your data aggregate, leaving at least one data partition as spare: storage aggregate create new_aggr_name -diskcount number_of_partitions - node active_node_name The data aggregate is created and is owned by the active node. Related information Clustered Data ONTAP 8.3 High-Availability Configuration Guide Adding disks to a node You add disks to a node to increase the number of hot spares, to add space to an aggregate, or to replace disks. Before you begin You must have confirmed that your platform model supports the type of disk you want to add. About this task You use this procedure to add physical disks to your node. If you are administering a node that uses virtual disks, for example, a platform based on Data ONTAP-v technology, see the installation and administration guide that came with your Data ONTAP-v system. Steps 1. Check the NetApp Support Site for newer disk and shelf firmware and Disk Qualification Package files. If your node does not have the latest versions, you must update them before installing the new disk. 2. Install the disks according to the hardware guide for your disk shelf or the hardware and service guide for your platform. The new disks are not recognized until they are assigned to a node and pool. You can assign the new disks manually, or you can wait for Data ONTAP to automatically assign the new disks if your node follows the rules for disk autoassignment. Managing disks using Data ONTAP | 41
3. After the new disks have all been recognized, verify their addition and their ownership information: storage aggregate show-spare-disks You should see the new disks, owned by the correct node and in the correct pool. 4. Optional: Zero the newly added disks: storage disk zerospares Disks that have been used previously in a Data ONTAP aggregate must be zeroed before they can be added to another aggregate. Zeroing the disks now can prevent delays in case you need to quickly increase the size of an aggregate. The disk zeroing command runs in the background and can take hours to complete, depending on the size of the non-zeroed disks in the node. Result The new disks are ready to be added to an aggregate, used to replace an existing disk, or placed onto the list of hot spares. Related concepts How automatic ownership assignment works for disks on page 56 Guidelines for assigning ownership for disks on page 58 Related information NetApp Hardware Universe NetApp Downloads: Disk Drive and Firmware When you need to update the Disk Qualification Package The Disk Qualification Package (DQP) adds full support for newly qualified drives. Before you update drive firmware or add new drive types or sizes to a cluster, you must update the DQP. A best practice is to also update the DQP regularly; for example, every quarter or semi-annually. You need to download and install the DQP in the following situations: • Whenever you add a new drive type or size to the node For example, if you already have 1-TB drives and add 2-TB drives, you need to check for the latest DQP update. • Whenever you update the disk firmware • Whenever newer disk firmware or DQP files are available • Whenever you upgrade to a new version of Data ONTAP. The DQP is not updated as part of a Data ONTAP upgrade. 42 | Physical Storage Management Guide
Related information NetApp Downloads: Disk Qualification Package NetApp Downloads: Disk Drive and Firmware Replacing disks that are currently being used in an aggregate You replace disks that are currently being used in an aggregate to swap out mismatched disks from a RAID group. Keeping your RAID groups homogeneous helps to optimize storage system performance. Before you begin You should already have an appropriate hot spare disk of the correct type, size, speed, and checksum type installed in your storage system. This spare disk must be assigned to the same system and pool as the disk it will replace. If the disk to be replaced is partitioned or in a storage pool, the spare disk (or its partitions, if applicable) can be owned by either node in the HA pair. For multi-disk carrier disks, you should have at least two hot spare disks available, to enable Data ONTAP to provide an optimal disk layout. About this task If you replace a smaller disk with a larger disk, the capacity of the larger disk is downsized to match that of the smaller disk; the usable capacity of the aggregate is not increased. Step 1. Replace the disk: storage disk replace -disk old_disk_name -replacement new_spare_disk_name -action start If you need to stop the disk replace operation, you can use the -action stop parameter. Result The old disk is converted to a spare disk, and the new disk is now used in the aggregate. Managing disks using Data ONTAP | 43
Replacing a self-encrypting disk Replacing a self-encrypting disk (SED) is similar to replacing a regular disk, except that there are some extra steps you must take to reenable Storage Encryption after you replace the disk. Before you begin You should know the key used by the SEDs on your storage system so that you can configure the replacement SED to use the same key. Steps 1. Ensure that reconstruction has started by entering the following command: aggr status -r The status of the disk should display as "Reconstructing". 2. Remove the failed disk and replace it with a new SED, following the instructions in the hardware guide for your disk shelf model. 3. Assign ownership of the newly replaced SED by entering the following command: disk assign disk_name 4. Confirm that the new disk has been properly assigned by entering the following command: disk encrypt show You should see the newly added disk in the output. 5. Encrypt the disk by entering the following command: disk encrypt rekey key_id disk_name 6. Finalize the replacement process by entering the following command: disk encrypt lock disk_name The newly replaced SED is ready for use, and Storage Encryption is enabled and working on this system. Converting a data disk to a hot spare Data disks can be converted to hot spares by destroying the aggregate that contains them. Before you begin The aggregate to be destroyed cannot contain volumes. 44 | Physical Storage Management Guide
About this task Converting a data disk to a hot spare does not change the ownership information for that disk. You must remove ownership information from a disk before moving it to another storage system. Step 1. Destroy the aggregate that contains the disk by entering the following command: storage aggregate delete -aggregate aggr_name All disks in use by that aggregate are converted to hot spare disks. Removing disks from a node How you remove a disk from a node depends how the disk is being used. By using the correct procedure, you can prevent unwanted AutoSupport notifications from being generated and ensure that the disk functions correctly if it is reused in another node. You cannot reduce the number of disks in an aggregate by removing data disks. The only way to reduce the number of data disks in an aggregate is to copy the data and transfer it to a new aggregate that has fewer data disks. If you are removing a partitioned drive, you must account for how both partitions are being used, and unpartition the drive before you remove it. If you are removing a disk in a multi-disk carrier, you must take special precautions. Related concepts Understanding root-data partitioning on page 30 How to determine when it is safe to remove a multi-disk carrier on page 28 Removing a failed disk A disk that is completely failed is no longer counted by Data ONTAP as a usable disk, and you can immediately disconnect the disk from the disk shelf. However, you should leave a partially failed disk connected long enough for the Rapid RAID Recovery process to complete. About this task If you are removing a disk because it has failed or because it is producing excessive error messages, you should not use the disk again in this or any other storage system. Steps 1. Find the disk ID of the failed disk by entering the following command: storage disk show -broken Managing disks using Data ONTAP | 45
If the disk does not appear in the list of failed disks, it might be partially failed, with a Rapid RAID Recovery in process. In this case, you should wait until the disk is present in the list of failed disks (which means that the Rapid RAID Recovery process is complete) before removing the disk. 2. Determine the physical location of the disk you want to remove by entering the following command: storage disk set-led -disk disk_name 2 The fault LED on the face of the disk is lit for 2 minutes. 3. Remove the disk from the disk shelf, following the instructions in the hardware guide for your disk shelf model. Removing a hot spare disk Removing a hot spare disk requires you to remove ownership information from the disk. This prevents the disk from causing problems when it is inserted into another node, and notifies Data ONTAP that you are removing the disk, which prevents unwanted AutoSupport messages. About this task Removing a hot spare disk does not make the contents of that disk inaccessible. If you need absolute assurance that the data contained by this disk is irretrievable, you should sanitize the disk instead of completing this procedure. Steps 1. Find the disk name of the hot spare disk you want to remove: storage aggregate show-spare-disks 2. Determine the physical location of the disk you want to remove: storage disk set-led -disk disk_name The fault LED on the face of the disk is lit. 3. If disk ownership automatic assignment is on, turn it off: storage disk option modify -node node_name -autoassign off 4. If disk ownership automatic assignment is on for the node's HA partner, turn it off on the partner node also. 5. Remove the software ownership information from the disk: storage disk removeowner disk_name 6. Remove the disk from the disk shelf, following the instructions in the hardware guide for your disk shelf model. 46 | Physical Storage Management Guide
7. If you turned off disk ownership automatic assignment previously, turn it on now: storage disk option modify -node node_name -autoassign on 8. If you turned off disk ownership automatic assignment previously on the node's HA partner, turn it on for the partner node also. Related concepts How to determine when it is safe to remove a multi-disk carrier on page 28 Related tasks Using disk sanitization to remove data from disks on page 48 Removing a data disk The only time that you should remove a data disk from a storage system is if the disk is not functioning correctly. If you want to remove a data disk so that it can be used in another system, you must convert it to a hot spare disk first. About this task You can cause Data ONTAP to fail the disk immediately or allow a disk copy to finish before the disk is failed. If you do not fail the disk immediately, you must wait for the disk copy to finish before physically removing the disk. This operation might take several hours, depending on the size of the disk and the load on the storage system. Do not immediately fail a disk unless it is causing immediate performance or availability issues for your storage system. Depending on your storage system configuration, additional disk failures could result in data loss. Steps 1. Determine the name of the disk you want to remove. If the disk is reporting errors, you can find the disk name in the log messages that report disk errors. The name is prefixed with the word “Disk”. 2. Determine the physical location of the disk you want to remove by entering the following command: storage disk set-led -disk disk_name 2 The red LED on the face of the disk is lit for 2 minutes. 3. Take the appropriate action based on whether you need to fail the disk immediately. Managing disks using Data ONTAP | 47
If you... Then... Can wait for the copy operation to finish (recommended) Enter the following command to pre-fail the disk: storage disk fail disk_name Data ONTAP pre-fails the specified disk and attempts to create a replacement disk by copying the contents of the pre-failed disk to a spare disk. If the copy operation is successful, then Data ONTAP fails the disk and the new replacement disk takes its place. If the copy operation fails, the pre-failed disk fails and the storage system operates in degraded mode until the RAID system reconstructs a replacement disk. Need to remove the disk immediately Enter the following command to cause the disk to fail immediately: storage disk fail -disk disk_name -immediate The disk fails and the storage system operates in degraded mode until the RAID system reconstructs a replacement disk. 4. Ensure that the disk you want to remove is shown as failed by entering the following command, and looking for its disk name: storage disk show -broken Do not remove the disk until its name appears in the list of failed disks. 5. Remove the failed disk from the disk shelf, following the instructions in the hardware guide for your disk shelf model. Related concepts About degraded mode on page 119 How to determine when it is safe to remove a multi-disk carrier on page 28 Using disk sanitization to remove data from disks Disk sanitization enables you to remove data from a disk or set of disks so that the data can never be recovered. Before you begin • The disks must be spare disks; they must be owned by a node, but not used in an aggregate. (If the disk is partitioned, neither partition can be in use in an aggregate.) • The disks cannot be part of a storage pool. 48 | Physical Storage Management Guide
About this task When disk sanitization is enabled, it disables some Data ONTAP commands. After disk sanitization is enabled on a node, it cannot be disabled. If you need to remove data from disks using Storage Encryption, do not use this procedure. Use the procedure for destroying data on disks using Storage Encryption. Steps 1. Enter the nodeshell for the node that owns the disks you want to sanitize: system node run -node node_name 2. Enable disk sanitization: options licensed_feature.disk_sanitization.enable on You are asked to confirm the command because it is irreversible. 3. If the disks you want to sanitize are partitioned, unpartition each disk: disk unpartition disk_name 4. Sanitize the specified disks: disk sanitize start [-p pattern1|-r [-p pattern2|-r [-p pattern3|-r]]] [-c cycle_count] disk_list Attention: Do not turn off power to the node, disrupt the storage connectivity, or remove target disks while sanitizing. If sanitizing is interrupted during the formatting phase, the formatting phase must be restarted and allowed to finish before the disks are sanitized and ready to be returned to the spare pool. If you need to abort the sanitization process, you can do so by using the disk sanitize abort command. If the specified disks are undergoing the formatting phase of sanitization, the abort does not occur until the phase is complete. -p pattern1 -p pattern2 -p pattern3 specifies a cycle of one to three user-defined hex byte overwrite patterns that can be applied in succession to the disks being sanitized. The default pattern is three passes, using 0x55 for the first pass, 0xaa for the second pass, and 0x3c for the third pass. -r replaces a patterned overwrite with a random overwrite for any or all of the passes. -c cycle_count specifies the number of times that the specified overwrite patterns are applied. The default value is one cycle. The maximum value is seven cycles. disk_list specifies a space-separated list of the IDs of the spare disks to be sanitized. 5. If you want to check the status of the disk sanitization process: disk sanitize status [disk_list] Managing disks using Data ONTAP | 49
6. After the sanitization process is complete, return the disks to spare status by entering the following command for each disk: disk sanitize release disk_name 7. Return to the clustered Data ONTAP CLI: exit 8. Determine whether all of the disks were returned to spare status: storage aggregate show-spare-disks If... Then... All of the sanitized disks are listed as spares You are done. The disks are sanitized and in spare status. Some of the sanitized disks are not listed as spares Complete the following steps: a. Enter advanced privilege mode: set -privilege advanced b. Assign the unassigned sanitized disks to the appropriate node by entering the following command for each disk: storage disk assign -disk disk_name -owner node_name c. Return the disks to spare status by entering the following command for each disk: storage disk unfail -disk disk_name -s -q d. Return to administrative mode: set -privilege admin Result The specified disks are sanitized and designated as hot spares. The serial numbers of the sanitized disks are written to /etc/log/sanitized_disks. Related concepts How disk sanitization works on page 19 50 | Physical Storage Management Guide
Stopping disk sanitization You can use the disk sanitize abort command to stop an ongoing sanitization process on one or more specified disks. Step 1. Enter the following command: disk sanitize abort disk_list This command is available through the nodeshell. If the specified disks are undergoing the disk formatting phase of sanitization, the process does not stop until the disk formatting is complete. Data ONTAP displays the message Sanitization abort initiated. After the process stops, Data ONTAP displays another message for each disk to inform you that sanitization is no longer in progress. Commands for managing disks You use the storage disk and storage aggregate commands to manage your disks. If you want to... Use this command... Display a list of spare disks, including partitioned disks, by owner storage aggregate show-spare-disks Display the disk RAID type, current usage, and RAID group by aggregate storage aggregate show-status Display the RAID type, current usage, aggregate, and RAID group, including spares, for physical disks storage disk show -raid Display a list of failed disks storage disk show -broken Illuminate the LED for a particular disk or shelf storage disk set-led Display the checksum type for a specific disk storage disk show -fields checksum- compatibility Display the checksum type for all spare disks storage disk show -fields checksum- compatibility -container-type spare Managing disks using Data ONTAP | 51
If you want to... Use this command... Display disk connectivity and placement information storage disk show -fields disk,primary-port,secondary- name,secondary-port,shelf,bay Display pre-cluster disk names for specific disks storage disk show -disk -fields diskpathnames Display a list of disks in the maintenance center storage disk show -maintenance Unpartition a disk system node run -node local - command disk unpartition Zero all non-zeroed disks storage disk zerospares Related information Clustered Data ONTAP 8.3 Commands: Manual Page Reference Commands for displaying information about storage shelves You use the storage shelf show command to display configuration and error information for your disk shelves. If you want to display... Use this command... General information about shelf configuration and hardware status storage shelf show Detailed information for a specific shelf, including stack ID storage shelf show -shelf Unresolved, customer actionable, errors by shelf storage shelf show -errors Bay information storage shelf show -bay Connectivity information storage shelf show -connectivity Cooling information, including temperature sensors and cooling fans storage shelf show -cooling Information about I/O modules storage shelf show -module Port information storage shelf show -port 52 | Physical Storage Management Guide
If you want to display... Use this command... Power information, including PSUs (power supply units), current sensors, and voltage sensors storage shelf show -power Related information Clustered Data ONTAP 8.3 Commands: Manual Page Reference Commands for displaying space usage information You use the storage aggregate and volume commands to see how space is being used in your aggregates and volumes and their Snapshot copies. To display information about... Use this command... Aggregates, including details about used and available space percentages, Snapshot reserve size, and other space usage information storage aggregate show storage aggregate show-space -snap- size-total,-used-including- snapshot-reserve How disks and RAID groups are used in an aggregate and RAID status storage aggregate show-status The amount of disk space that would be reclaimed if you deleted a specific Snapshot copy volume snapshot compute-reclaimable (advanced) The amount of space used by a volume volume show -fields size,used,available,percent-used volume show-space The amount of space used by a volume in the containing aggregate volume show-footprint Related information Clustered Data ONTAP 8.3 Commands: Manual Page Reference Managing disks using Data ONTAP | 53
Managing ownership for disks Disk ownership determines which node owns a disk and what pool a disk is associated with. Data ONTAP stores ownership information directly on the disk. Types of disk ownership The HA or Disaster Recovery (DR) state of the system that owns a disk can affect which system has access to the disk. This means that there are several types of ownership for disks. Disk ownership information is set either by Data ONTAP or by the administrator, and recorded on the disk, in the form of the controller module's unique system ID (obtained from a node's NVRAM card or NVMEM board). Disk ownership information displayed by Data ONTAP can take one or more of the following forms. Note that the names used vary slightly depending on the context. • Owner (or Current owner) This is the system that can currently access the disk. • Original owner (or Home owner) If the system is in HA takeover, then Owner is changed to the system that took over the node, and Original owner or Home owner reflects the system that owned the disk before the takeover. • DR home owner If the system is in a MetroCluster switchover, DR home owner reflects the value of the Home owner field before the switchover occurred. Reasons to assign ownership of disks and array LUNs Storage system ownership must be assigned for disks and array LUNs before they become an effective part of your system. You must explicitly assign ownership for array LUNs. Disks can be automatically or manually assigned. You assign ownership of a disk or array LUN to accomplish the following actions: • Associate the disk or array LUN with a specific storage system. For a stand-alone system, all disks and array LUNs are owned by that system. In an HA configuration, the disks and array LUNs can be owned by either system. • Enable the disk or array LUN to be used and managed by the system that owns it. Unowned disks cannot be used as spares and do not receive the automatic firmware updates that owned disks do. 54 | Physical Storage Management Guide
• Associate the disk or array LUN with a specific SyncMirror pool (when SyncMirror is in use). If SyncMirror is not in use, all disks and array LUNs are in pool0. How disks and array LUNs become available for use When you add a disk or array LUN to a system running Data ONTAP, the disk or array LUN goes through several stages before it can be used by Data ONTAP to store data or parity information. The process for making a disk available for use differs slightly from the process for making an array LUN available for use. Both processes are shown in the following diagram: Add to aggregate (optional) Install a new disk on a disk shelf Storage array Create array LUNs Make array LUNs available to Data ONTAP System running Data ONTAP Data ONTAP Manual assignment of array LUNs to a system running Data ONTAP Automatic or manual assignment of a new disk to a system running Data ONTAP Spare disk or array LUN It is owned by the storage system, but it cannot be used yet. In-use disk or array LUN The disk or LUN is in use by the system that owns it. Unowned disk or array LUN The process for disks includes the following actions: 1. The administrator physically installs the disk into a disk shelf. Data ONTAP can see the disk, but the disk is still unowned. 2. If the system is configured to support disk autoassignment, Data ONTAP assigns ownership for the disk; otherwise, the administrator must assign ownership of the disk manually. The disk is now a spare disk. 3. The administrator or Data ONTAP adds the disk to an aggregate. The disk is now in use by that aggregate. It could contain data or parity information. Managing ownership for disks | 55
The process for array LUNs includes the following actions: 1. The storage array administrator creates the array LUN and makes it available to Data ONTAP. Data ONTAP can see the array LUN, but the array LUN is still unowned. 2. The Data ONTAP administrator assigns ownership of the array LUN to a Data ONTAP system. The array LUN is now a spare array LUN. 3. The Data ONTAP administrator adds the array LUN to an aggregate. The array LUN is now in use by that aggregate and is storing data. How automatic ownership assignment works for disks If your configuration follows some basic rules to avoid ambiguity, Data ONTAP can automatically assign ownership and pool membership for disks. Automatic disk autoassignment uses the configured autoassignment policy to determine the correct ownership for disks. The autoassignment policy can be stack, shelf, bay, or default. In most cases, all disks in a stack (or loop) can be assigned to the same node and pool. For mid- and high-end platforms, the default autoassignment policy is equivalent to the stack policy. For entry-level platforms (FAS2xxx), the default autoassignment policy is equivalent to the bay policy, which supports two controllers sharing the same shelf. For this policy, disks in odd bay numbers are assigned together, and disks in even bay numbers are assigned together. For partitioned disks, autoassignment operates on the container disk and also both of the partitions. If you decide to change the way Data ONTAP has assigned the disks, you can do so at any time. You can also change autoassignment to use a different assignment policy, although doing so does not affect currently assigned disks. If you need to temporarily remove disk ownership for one or more disks while you perform an administrative task, you must disable automatic ownership assignment first to prevent Data ONTAP from immediately reassigning ownership for those disks. Automatic ownership assignment is not available for array LUNs or virtual disks. Which disk autoassignment policy to use Usually, you can use the default autoassignment policy, which is equivalent to the stack policy for most systems, and to the bay policy for entry level systems (FAS2xxx). However, for some configurations, you might need to change the autoassignment policy. Use the appropriate autoassignment, based on your configuration: If you are using... Then use the autoassignment policy value of... Stand-alone entry level system stack 56 | Physical Storage Management Guide
If you are using... Then use the autoassignment policy value of... Entry level systems in an HA configuration with a single, shared shelf bay Entry level systems in an HA configuration with one stack of two or more shelves shelf MetroCluster configurations with one stack per node, two or more shelves shelf All other configurations stack When automatic ownership assignment is invoked Automatic disk ownership assignment does not happen immediately after disks are introduced into the storage system. Automatic ownership assignment is invoked at the following times: • Every five minutes during normal system operation • Ten minutes after the initial system initialization This delay enables the person configuring the system enough time to finish the initial disk assignments so that the results of the automatic ownership assignment are correct. • Whenever you enable automatic ownership assignment. How disk ownership works for platforms based on Data ONTAP-v technology You manage ownership for virtual disks by using the same commands you use for physical disks. However, automatic ownership assignment works differently for virtual disks. Storage systems based on Data ONTAP-v technology, for example, Data ONTAP Edge systems, automatically assign ownership for all the virtual data disks you defined during the initial system setup. Automatic assignment is not available for any virtual disks you add after the initial system setup. For information about adding virtual disks and managing a storage system based on Data ONTAP-v technology, see the Data ONTAP Edge Installation and Administration Guide. Managing ownership for disks | 57
Guidelines for assigning ownership for disks When you assign ownership for disks, you need to follow certain guidelines to maximize fault isolation and to keep automatic ownership assignment working. The guidelines are impacted by your autoassignment policy. Use these guidelines when you assign ownership for disks: • Assign disks so that your autoassignment policy group is homogenous. For example, if your autoassignment policy is stack (the default for most platforms), assign all disks on the same stack or loop to the same node and pool. However, if your autoassignment policy is bay (the default for FAS2xxx models), keep your disks in even-numbered bays and odd- numbered bays homogenous by node and pool. • If your autoassignment policy is stack, assign all stacks connected to the same adapter to the same pool. • Assign disks in the same multi-disk carrier to the same node and pool. Assigning ownership for disks Disks must be owned by a node before they can be used in an aggregate. If your cluster is not configured to use automatic disk ownership assignment, you must assign ownership manually. About this task This procedure is only for disks that are not partitioned. You can use the wildcard character to assign more than one disk at once. If you are reassigning a spare disk that is already owned by a different node, you must use the - force option for the storage disk assign command. You cannot reassign a disk that is in use in an aggregate. Steps 1. Display all unowned disks: storage disk show -container-type unassigned 2. Assign each disk: storage disk assign -disk disk_name -owner owner_name 58 | Physical Storage Management Guide
Related concepts How automatic ownership assignment works for disks on page 56 Related tasks Assigning ownership for disks partitioned for root-data partitioning on page 59 Assigning ownership for disks partitioned for root-data partitioning For partitioned disks, there are three different entities for ownership: the container disk, the data partition, and the root partition. You can set the ownership of the container disk or the partitions manually or by using autoassignment—just as you do for unpartitioned disks. About this task You can think of the three owned entities (the container disk and the two partitions) as being collectively owned by the HA pair. The container disk and the two partitions do not all need to be owned by the same node in the HA pair as long as they are all owned by one of the nodes in the HA pair. However, when you use a partition in an aggregate, it must be owned by the same node that owns the aggregate. Steps 1. Display the current ownership for the partitioned disk: storage disk show -disk disk_name -partition-ownership 2. Enter the appropriate command, depending on which ownership entity you want to assign ownership for: If you want to assign ownership for the... Use this command... Container disk storage disk assign -disk disk_name -owner owner_name Data partition storage disk assign -disk disk_name -owner owner_name -data true Root partition storage disk assign -disk disk_name -owner owner_name -root true If any of the ownership entities are already owned, then you must include the -force option. Related concepts How automatic ownership assignment works for disks on page 56 Managing ownership for disks | 59
Related tasks Assigning ownership for disks on page 58 Removing ownership from a disk Data ONTAP writes disk ownership information to the disk. Before you remove a spare disk or its shelf from a node, you should remove its ownership information so that it can be properly integrated into another node. Before you begin The disk you want to remove ownership from must meet the following requirements: • It must be a spare disk. You cannot remove ownership from a disk that is being used in an aggregate. • It cannot be in the maintenance center. • It cannot be undergoing sanitization. • It cannot be failed. It is not necessary to remove ownership from a failed disk. About this task If you have automatic disk assignment enabled, Data ONTAP could automatically reassign ownership before you remove the disk from the node. For this reason, you disable automatic ownership assignment until the disk is removed, and then reenable it. Steps 1. If disk ownership automatic assignment is on, turn it off: storage disk option modify -node node_name -autoassign off 2. If needed, repeat the previous step for the node's HA partner. 3. Remove the software ownership information from the disk: storage disk removeowner disk_name To remove ownership information from multiple disks, use a comma-separated list. Example storage disk removeowner sys1:0a.23,sys1:0a.24,sys1:0a.25 4. If the disk is partitioned for root-data partitioning, remove ownership from the partitions by entering both of the following commands: 60 | Physical Storage Management Guide
storage disk removeowner disk_name -root true storage disk removeowner disk_name -data true Both partitions are no longer owned by any node. 5. If you turned off disk ownership automatic assignment previously, turn it on after the disk has been removed or reassigned: storage disk option modify -node node_name -autoassign on 6. If needed, repeat the previous step for the node's HA partner. Configuring automatic ownership assignment of disks You can configure Data ONTAP to automatically assign disk ownership according to a disk's stack, shelf, or bay. Before you begin • Your system must adhere to the requirements for automatic disk ownership. • If you have multiple stacks or shelves that must have different ownership, one disk must have been manually assigned on each stack or shelf so that automatic ownership assignment will work on each stack or shelf. About this task The behavior of the default autoassignment policy depends on the system model. For entry level models, the default policy is equivalent to the bay policy. For all other systems, it is equivalent to the stack policy. Steps 1. The action you take depends on whether you want to set up automatic ownership assignment at the stack (or loop), shelf, or bay level: If you want to... Then use the following command... Configure automatic ownership assignment at the stack or loop level storage disk option modify -autoassign-policy stack Configure automatic ownership assignment at the shelf level storage disk option modify -autoassign-policy shelf Managing ownership for disks | 61
If you want to... Then use the following command... Configure automatic ownership assignment at the bay level storage disk option modify -autoassign-policy bay Turn off automatic ownership assignment storage disk option modify -autoassign off 2. Verify the automatic assignment settings for the disks: storage disk option show Example cluster1::> storage disk option show Node BKg. FW. Upd. Auto Copy Auto Assign Auto Assign Policy ------------- ------------- ------------ ------------- ------------------ cluster1-1 on on on default cluster1-2 on on on default Related concepts How automatic ownership assignment works for disks on page 56 Which disk autoassignment policy to use on page 56 When automatic ownership assignment is invoked on page 57 How you use the wildcard character with the disk ownership commands You can use the wildcard character (*) with some commands, including commands to manage disk ownership. However, you should understand how Data ONTAP expands the wildcard character to ensure that you operate on the correct set of disks. You can use the wildcard character with the following disk ownership commands: • storage disk assign • storage disk show • storage disk removeowner When you use the wildcard character with these commands, Data ONTAP expands it with zero or more characters to create a list of disk names that will be operated on by the command. This can be 62 | Physical Storage Management Guide
very useful, for example, when you want to assign all of the disks attached to a particular port or switch. Note: Be careful when you use the wildcard character. It is accepted anywhere in the disk name string, and is a simple string substitution. Therefore, you might get unexpected results. For example, to operate on all disks on shelf 1 of stack 1, you would use the following syntax: 1.1.* However, if you left off the second “.”, as in the following snippet, you would operate on all disks in stacks 1, 10, 11, and so on: 1.1* Managing ownership for disks | 63
Managing array LUNs using Data ONTAP For Data ONTAP to be able to use storage on a storage array, some tasks must be done on the storage array and some tasks must be done in Data ONTAP. For example, the storage array administrator must create array LUNs for Data ONTAP use and map them to Data ONTAP. You can then assign them to nodes running Data ONTAP. If the storage array administrator wants to make configuration changes to an array LUN after it is assigned to a node, for example to resize it, you might need to perform some activities in Data ONTAP before it is possible to reconfigure the LUN on the storage array. Related concepts How ownership for disks and array LUNs works on page 66 Data ONTAP systems that can use array LUNs on storage arrays V-Series (“V”) systems and new FAS platforms released in Data ONTAP 8.2.1 and later can use array LUNs if the proper license is installed. In discussions in the Data ONTAP and FlexArray Virtualization documentation, these systems are collectively referred to as Data ONTAP systems when it is necessary to make it clear which information applies to them and what information applies to storage arrays. Note: Starting with Data ONTAP 8.2.1, the capability of using LUNs on a storage array, formerly identified as V-Series functionality, has a new name—Data ONTAP FlexArray Virtualization Software. The capability of using array LUNs continues to be available as a licensed feature in Data ONTAP. Systems prior to Data ONTAP 8.2.1 that can use array LUNs The only systems released prior to Data ONTAP 8.2.1 that can use array LUNs are V-Series systems —systems with a “V” or “GF” prefix. A V-Series system is an open storage controller that virtualizes storage from storage array vendors, native disks, or both into a single heterogeneous storage pool. Note: Almost all Data ONTAP platforms released prior to Data ONTAP 8.2.1 were released with FAS and V-Series equivalent models (for example, a FAS6280 and a V6280 ). (For a few systems, there were no “V” equivalent models.) Although both types of models could access native disks, only the V-Series systems (a “V” or “GF” prefix) could attach to storage arrays. 64 | Physical Storage Management Guide
Systems in Data ONTAP 8.2.1 and later that can use array LUNs Starting with Data ONTAP 8.2.1, the model for how platforms are released and the storage they can use changes. Attaching to storage arrays is no longer limited to V-Series systems. Starting with Data ONTAP 8.2.1, all new platforms are released as a single hardware model. This single hardware model has a FAS prefix; there are no longer separate “V” and FAS models for new platforms. If the V_StorageAttach license package is installed on a new FAS model, it can attach to storage arrays. (This is the same license required on a V-Series system.) Important: FAS systems released prior to Data ONTAP 8.2.1 cannot use LUNs on storage arrays, even if they are upgraded to Data ONTAP 8.2.1 or later; only the “V” equivalent of a platform can use array LUNs. Installing the license for using array LUNs The V_StorageAttach license must be installed on each Data ONTAP node that you want to use with array LUNs. It is not a single license for the cluster. Array LUNs cannot be used in aggregates until a license is installed. Before you begin • The cluster must be installed. • You must have the license key for the V_StorageAttach license. NetApp Support About this task You need not perform this procedure if the license key for the V_StorageAttach package is already installed. If the Data ONTAP system is ordered with disks, the factory typically installs the license package for you. Alternatively, many customers install all necessary licenses early in the installation process. Steps 1. For each Data ONTAP node in the cluster for use with array LUNs, enter the following command on the node: system license add license key Managing array LUNs using Data ONTAP | 65
Example vgv3170f41a> license Serial Number: nnnnnnnn Owner: mysystem1a Package Type Description Expiration ----------------- ------- --------------------- -------------------- V_StorageAttach license Virtual Attached Storage 2. Look at the output to confirm that the V_StorageAttach package is shown. How ownership for disks and array LUNs works Disk and array LUN ownership determines which node owns a disk or array LUN and what pool a disk or array LUN is associated with. Understanding how ownership works enables you to maximize storage redundancy and manage your hot spares effectively. Data ONTAP stores ownership information directly on the disk or array LUN. Reasons to assign ownership of disks and array LUNs Storage system ownership must be assigned for disks and array LUNs before they become an effective part of your system. You must explicitly assign ownership for array LUNs. Disks can be automatically or manually assigned. You assign ownership of a disk or array LUN to accomplish the following actions: • Associate the disk or array LUN with a specific storage system. For a stand-alone system, all disks and array LUNs are owned by that system. In an HA configuration, the disks and array LUNs can be owned by either system. • Enable the disk or array LUN to be used and managed by the system that owns it. Unowned disks cannot be used as spares and do not receive the automatic firmware updates that owned disks do. • Associate the disk or array LUN with a specific SyncMirror pool (when SyncMirror is in use). If SyncMirror is not in use, all disks and array LUNs are in pool0. How disks and array LUNs become available for use When you add a disk or array LUN to a system running Data ONTAP, the disk or array LUN goes through several stages before it can be used by Data ONTAP to store data or parity information. The process for making a disk available for use differs slightly from the process for making an array LUN available for use. Both processes are shown in the following diagram: 66 | Physical Storage Management Guide
Add to aggregate (optional) Install a new disk on a disk shelf Storage array Create array LUNs Make array LUNs available to Data ONTAP System running Data ONTAP Data ONTAP Manual assignment of array LUNs to a system running Data ONTAP Automatic or manual assignment of a new disk to a system running Data ONTAP Spare disk or array LUN It is owned by the storage system, but it cannot be used yet. In-use disk or array LUN The disk or LUN is in use by the system that owns it. Unowned disk or array LUN The process for disks includes the following actions: 1. The administrator physically installs the disk into a disk shelf. Data ONTAP can see the disk, but the disk is still unowned. 2. If the system is configured to support disk autoassignment, Data ONTAP assigns ownership for the disk; otherwise, the administrator must assign ownership of the disk manually. The disk is now a spare disk. 3. The administrator or Data ONTAP adds the disk to an aggregate. The disk is now in use by that aggregate. It could contain data or parity information. The process for array LUNs includes the following actions: 1. The storage array administrator creates the array LUN and makes it available to Data ONTAP. Data ONTAP can see the array LUN, but the array LUN is still unowned. 2. The Data ONTAP administrator assigns ownership of the array LUN to a Data ONTAP system. The array LUN is now a spare array LUN. 3. The Data ONTAP administrator adds the array LUN to an aggregate. The array LUN is now in use by that aggregate and is storing data. Managing array LUNs using Data ONTAP | 67
What it means for Data ONTAP to own an array LUN Data ONTAP cannot use an array LUN presented to it by a storage array until you configure a logical relationship in Data ONTAP that identifies a specific system running Data ONTAP as the owner of the array LUN. A storage array administrator creates array LUNs and makes them available to specified FC initiator ports of storage systems running Data ONTAP. (The process for how to do this varies among storage array vendors.) When you assign an array LUN to a system running Data ONTAP, Data ONTAP writes data to the array LUN to identify that system as the owner of the array LUN. Thereafter, Data ONTAP ensures that only the owner can write data to and read data from the array LUN. From the perspective of Data ONTAP, this logical relationship is referred to as disk ownership because Data ONTAP considers an array LUN to be a virtual disk. From the perspective of Data ONTAP, you are assigning disks to a storage system. An advantage of the disk ownership scheme is that you can make changes through the Data ONTAP software that, on typical hosts, must be done by reconfiguring hardware or LUN access controls. For example, through Data ONTAP you can balance the load of requests among a group of systems running Data ONTAP by moving data service from one system to another, and the process is transparent to most users. You do not need to reconfigure hardware or the LUN access controls on the storage array to change which system running Data ONTAP is the owner and, therefore, servicing data requests. Attention: The Data ONTAP software-based scheme provides ownership control only for storage systems running Data ONTAP; it does not prevent a different type of host from overwriting data in an array LUN owned by a system running Data ONTAP. Therefore, if multiple hosts are accessing array LUNs through the same storage array port, be sure to use LUN security on your storage array to prevent the systems from overwriting each other's array LUNs. Array LUN reconfiguration, such as resizing the array LUN, must be done from the storage array. Before such activities can occur, you must release Data ONTAP ownership of the array LUN. Why you might assign array LUN ownership after installation For a Data ONTAP system ordered with disk shelves, you are not required to set up the system to work with array LUNs during initial installation. In the case of a Data ONTAP system using only array LUNs, you need to assign only two array LUNs during initial installation. If you ordered your Data ONTAP system with disk shelves, you do not need to assign any array LUNs initially because the factory installs the root volume on a disk for you. If you are using only array LUNs, you must configure one array LUN for the root volume and one array LUN as a spare for core dumps during initial installation. In either case, you can assign ownership of additional array LUNs to your system at any time after initial installation. After you configure your system, you might assign ownership of an array LUN in the following circumstances: 68 | Physical Storage Management Guide
• You ordered your Data ONTAP system with native disk shelves and did not set up your system to work with array LUNs initially • You left some LUNs that the storage array presented to Data ONTAP unowned, and you now need to use the storage • Another system released the ownership of a particular array LUN and you want this system to use the LUN • The storage array administrator has not made the LUNs available to Data ONTAP during the initial system configuration, and you want to use the storage. Examples showing when Data ONTAP can use array LUNs After an array LUN has been assigned to a storage system, it can be added to an aggregate and used for storage or it can remain a spare LUN until it is needed for storage. No storage system owns the LUNs yet In this example, the storage array administrator made the array LUNs available to Data ONTAP. However, system vs1 has not yet been configured to “own” any of the LUNs. Therefore, it cannot read data from or write data to any array LUNs on the storage array: Only some array LUNs are owned In this example, vs1 was configured to own array LUNs 1 and 2, but not array LUNs 3 and 4. LUNs 3 and 4 are still available to Data ONTAP, however, and can be assigned to a storage system later: Data ONTAP used LUN 1 for the root volume. System vs1 can read data from and write data to LUN 1, because LUN 1 is in an aggregate. LUN 2 remains a spare LUN because it has not yet been added to an aggregate. System vs1 cannot read data from and write data to LUN 2 while it is a spare. Managing array LUNs using Data ONTAP | 69
After you perform initial setup of the storage system, you could configure vs1 to also own LUN 3, LUN 4, both, or neither, depending on your storage needs. Ownership of LUNs in an HA pair In this example, two storage systems running Data ONTAP are configured in an HA pair. In an HA pair, only one node can be the owner of a particular LUN, but both nodes must be able to see the same LUNs so that the partner can take over if the owning node becomes unavailable. LUN 1 through LUN 4 were created on the storage array and mapped to the ports on the storage array to which the storage systems are connected. All four LUNs are visible to each node in the HA pair. Assume that during initial setup vs1 was assigned ownership of LUN 1 and LUN 2. LUN 1 was automatically added to the root volume, so LUN 1 is now “in use” by vs1. LUN 2 remains a spare until it is explicitly added to an aggregate on vs1. Similarly, assume that during initial setup vs2 was assigned ownership of LUN 3 and LUN 4, with LUN 3 assigned to the root volume. LUN 4 remains a spare LUN until it is explicitly added to an aggregate. The key points of this example are as follows: 70 | Physical Storage Management Guide
• By deploying the storage systems in an HA pair, one system can take over services for its partner if the partner becomes unavailable. • Only one storage system can own a specific array LUN. However, all array LUNs assigned to a node in an HA pair must be visible to—but not assigned to or owned by—the other node in the HA pair. • By deploying two switches, if one switch fails, the other switch provides the alternate path to the storage array. • Both switches must be zoned correctly so that each storage system in the HA pair can see the array LUNs owned by its partner. Assigning ownership of array LUNs Array LUNs must be owned by a node before they can be added to an aggregate to be used as storage. Before you begin • Back-end configuration testing (testing of the connectivity and configuration of devices behind the Data ONTAP systems) must be completed. • Array LUNs that you want to assign must be presented to the Data ONTAP systems. About this task You can assign ownership of array LUNs that have the following characteristics: • They are unowned. • They have no storage array configuration errors, such as the following: ◦ The array LUN is smaller than or larger than the size that Data ONTAP supports. ◦ The LDEV is mapped on only one port. ◦ The LDEV has inconsistent LUN IDs assigned to it. ◦ The LUN is available on only one path. Data ONTAP issues an error message if you try to assign ownership of an array LUN with back-end configuration errors that would interfere with the Data ONTAP system and the storage array operating together. You must fix such errors before you can proceed with array LUN assignment. Data ONTAP alerts you if you try to assign an array LUN with a redundancy error: for example, all paths to this array LUN are connected to the same controller or only one path to the array LUN. You can fix a redundancy error before or after assigning ownership of the LUN. Managing array LUNs using Data ONTAP | 71
Steps 1. Enter the following command to see the array LUNs that have not yet been assigned to a node: storage disk show -container-type unassigned 2. Enter the following command to assign an array LUN to this node: storage disk assign -disk arrayLUNname -owner nodename If you want to fix a redundancy error after disk assignment instead of before, you must use the - force parameter with the storage disk assign command. Related concepts How ownership for disks and array LUNs works on page 66 Related tasks Modifying assignment of spare array LUNs on page 73 Verifying the back-end configuration It is important to detect and resolve any configuration errors before you bring online the Data ONTAP configuration with array LUNs in a production environment. You start installation verification by using storage array config show command. The storage array config show command shows how storage arrays connect to the cluster. If Data ONTAP detects an error in the back-end configuration, the following message is displayed at the bottom of the storage array config show output: Warning: Configuration errors were detected. Use 'storage errors show' for detailed information. You then use the storage errors show output to see details of the problem, at the LUN level. You must fix any errors shown by the storage errors show command. Related information FlexArray Virtualization Implementation Guide for Third-Party Storage 72 | Physical Storage Management Guide
Modifying assignment of spare array LUNs You can change the ownership of a spare array LUN to another node. You might want to do this for load balancing over the nodes. Steps 1. At the console of the node that owns the array LUN you want to reassign, enter the following command to see a list of spare array LUNs on the node: storage disk show -owner local The array LUNs owned by the node, both spares and LUNs in aggregates, are listed. 2. Confirm that the LUN you want to reassign to another node is a spare LUN. 3. Enter the following command to assign ownership of the array LUN to another node: storage disk assign arrayLUNname -owner new_owner_name -force Note: The array LUN ownership is not changed if the -force option is not used or if the array LUN was already added to an aggregate. 4. Enter the following command to verify that the ownership of the spare array LUN was changed to the other node: storage disk show -owner local The spare array LUN that you changed to the new owner should no longer appear in the list of spares. If the array LUN still appears, repeat the command to change ownership. 5. On the destination node, enter the following command to verify that the spare array LUN whose ownership you changed is listed as a spare owned by the destination node: storage disk show -owner local After you finish You must add the array LUN to an aggregate before it can be used for storage. Related concepts How ownership for disks and array LUNs works on page 66 Related tasks Assigning ownership of array LUNs on page 71 Managing array LUNs using Data ONTAP | 73
Array LUN name format In Data ONTAP 8.3, the name assigned to an array LUN has a new format to ensure that the name is unique across a cluster. The array LUN name consists of two components and looks like the following: <array_prefix>.<offset>, for example EMC-1.1. • The array_prefix is a unique prefix that Data ONTAP assigns by default to each storage array. This field is composed of <array_name-array_instance> (EMC-1 in this case). array_name can be denoted by the first three letters of the name of the vendor. If there is more than one array from the same vendor, the value of array_instance proceeds in an ascending order. • The offset is the ascending virtual disk number that Data ONTAP assigns to each LUN. It is independent of the LUN ID of the host. You can modify the <array_prefix> field by using the storage array modify -name - prefix command. Pre-cluster array LUN name format Before a node joins to a cluster or when the system is in the maintenance mode, the array LUN name follows a format used before Data ONTAP 8.3, the pre-cluster format. In this format, the array LUN name is a path-based name that includes the devices in the path between the Data ONTAP system and the storage array, ports used, and the SCSI LUN ID on the path that the storage array presents externally for mapping to hosts. On a Data ONTAP system that supports array LUNs, each array LUN can have multiple names because there are multiple paths to each LUN. 74 | Physical Storage Management Guide
Array LUN name format for Data ONTAP systems Configuration Array LUN name format Component descriptions Direct-attached node- name.adapter.idlun-id node-name is the name of the clustered node. With clustered Data ONTAP, the node name is prepended to the LUN name so that the path-based name is unique within the cluster. adapter is the adapter number on the Data ONTAP system. id is the channel adapter port on the storage array. lun-id is the array LUN number that the storage array presents to hosts. Example: node1.0a.0L1 Fabric-attached node-name:switch- name:port.idlun-id node-name is the name of the node. With clustered Data ONTAP, the node name is prepended to the LUN name so that the path-based name is unique within the cluster. switch-name is the name of the switch. port is the switch port that is connected to the target port (the end point). id is the device ID. lun-id is the array LUN number that the storage array presents to hosts. Example: node1:brocade3:6.126L1 Related concepts Pre-cluster drive name formats on page 15 Managing array LUNs using Data ONTAP | 75
Checking the checksum type of spare array LUNs If you plan to add a spare array LUN to an aggregate by specifying its name, you need to make sure that the checksum type of the array LUN you want to add is the same as the aggregate checksum type. About this task You cannot mix array LUNs of different checksum types in an array LUN aggregate. The checksum type of the aggregate and the checksum type of the array LUNs added to it must be the same. If you specify a number of spare array LUNs to be added to an aggregate, by default Data ONTAP selects array LUNs of the same checksum type as the aggregate. Note: The checksum type of all newly created aggregates using zoned checksum array LUNS is advanced zoned checksum (AZCS). Zoned checksum type continues to be supported for existing zoned aggregates. Zoned checksum spare array LUNs added to an existing zoned checksum aggregate continue to be zoned checksum array LUNs. Zoned checksum spare array LUNs added to an AZCS checksum type aggregate use the AZCS checksum scheme for managing checksums. Step 1. Check the checksum type of the spare array LUNs by entering the following command: storage disk show -fields checksum-compatibility -container-type spare You can add a block checksum array LUN to a block checksum aggregate and a zoned array LUN to an advanced zoned checksum (AZCS) aggregate. Related tasks Changing the checksum type of an array LUN on page 76 Changing the checksum type of an array LUN You must change the checksum type of an array LUN if you want to add it to an aggregate that has a different checksum type than the checksum type of the LUN. Before you begin You must have reviewed the tradeoffs between performance in certain types of workloads and storage capacity utilization of each checksum type. The FlexArray Virtualization Installation Requirements and Reference Guide contains information about checksum use for array LUNs. You can also contact your Sales Engineer for details about using checksums. 76 | Physical Storage Management Guide
About this task • You must assign a zoned checksum type to an array LUN that you plan to add to an advanced zoned checksum (AZCS) aggregate. When a zoned checksum array LUN is added to an AZCS aggregate, it becomes an advanced zoned checksum array LUN. Similarly, when a zoned checksum array LUN is added to a zoned aggregate, it is a zoned checksum type. • You cannot modify the checksum of array LUNs while assigning ownership. You can modify the checksum only on already assigned array LUNs. Step 1. Enter the following command to change the checksum type: storage disk assign -disk disk name -owner owner -c new_checksum_type disk name is the array LUN whose checksum type you want to change. owner is the node to which the array LUN is assigned. new_checksum_type can be block or zoned. Example storage disk assign -disk EMC-1.1 -owner system147b -c block The checksum type of the array LUN is changed to the new checksum type you specified. Related tasks Checking the checksum type of spare array LUNs on page 76 Prerequisites to reconfiguring an array LUN on the storage array If an array LUN has already been assigned (through Data ONTAP) to a particular Data ONTAP system, the information that Data ONTAP wrote to the array LUN must be removed before the storage administrator attempts to reconfigure the array LUN on the storage array. When the storage array presents an array LUN to Data ONTAP, Data ONTAP collects information about the array LUN (for example, its size) and writes that information to the array LUN. Data ONTAP cannot dynamically update information that it wrote to an array LUN. Therefore, before the storage array administrator reconfigures an array LUN, you must use Data ONTAP to change the state of the array LUN to unused. The array LUN is unused from the perspective of Data ONTAP. While changing the state of the array LUN to unused, Data ONTAP does the following: • Terminates I/O operations to the array LUN Managing array LUNs using Data ONTAP | 77
• Removes the label for RAID configuration information and the persistent reservations from the array LUN, which makes the array LUN unowned by any Data ONTAP system After this process finishes, no Data ONTAP information remains in the array LUN. You can do the following after the array LUN's state is changed to unused: • Remove the mapping of the array LUN to Data ONTAP and make the array LUN available to other hosts. • Resize the array LUN or change its composition. If you want Data ONTAP to resume using the array LUN after its size or composition is changed, you must present the array LUN to Data ONTAP again, and assign the array LUN to a Data ONTAP system again. Data ONTAP is aware of the new array LUN size or composition. Related tasks Changing array LUN size or composition on page 78 Changing array LUN size or composition Reconfiguring the size or composition of an array LUN must be done on the storage array. If an array LUN has already been assigned to a Data ONTAP system, you must use Data ONTAP to change the state of the array LUN to unused before the storage array administrator can reconfigure it. Before you begin The array LUN must be a spare array LUN before you can change its state to unused. Steps 1. On the Data ONTAP system, enter the following command to remove ownership information: storage disk removeowner -disk arrayLUNname 2. On the storage array, complete the following steps: a. Unmap (unpresent) the array LUN from the Data ONTAP systems so that they can no longer see the array LUN. b. Change the size or composition of the array LUN. c. If you want Data ONTAP to use the array LUN again, present the array LUN to the Data ONTAP systems again. At this point, the array LUN is visible to the FC initiator ports to which the array LUN was presented, but it cannot be used by any Data ONTAP systems yet. 78 | Physical Storage Management Guide
3. Enter the following command on the Data ONTAP system that you want to be the owner of the array LUN: storage disk assign -disk arrayLUNname -owner nodename After the ownership information is removed, the array LUN cannot be used by any Data ONTAP system until the array LUN is assigned again to a system. You can leave the array LUN as a spare or add it to an aggregate. You must add the array LUN to an aggregate before the array LUN can be used for storage. Related concepts Prerequisites to reconfiguring an array LUN on the storage array on page 77 Removing one array LUN from use by Data ONTAP If the storage array administrator no longer wants to use a particular array LUN for Data ONTAP, you must remove the information that Data ONTAP wrote to the LUN (for example, size and ownership) before the administrator can reconfigure the LUN for use by another host. Before you begin If the LUN that the storage array administrator no longer wants Data ONTAP to use is in an aggregate, you must take the aggregate offline and destroy the aggregate before starting this procedure. Taking an aggregate offline and destroying it changes the data LUN to a spare LUN. Step 1. Enter the following command: storage disk removeowner -disk LUN_name LUN_name is the name of the array LUN. Preparing array LUNs before removing a Data ONTAP system from service You must release the persistent reservations on all array LUNs assigned to a Data ONTAP system before removing the system from service. About this task When you assign Data ONTAP ownership of an array LUN, Data ONTAP places persistent reservations (ownership locks) on that array LUN to identify which Data ONTAP system owns the LUN. If you want the array LUNs to be available for use by other types of hosts, you must remove the persistent reservations that Data ONTAP put on those array LUNs; some arrays do not allow you Managing array LUNs using Data ONTAP | 79
to destroy a reserved LUN if you do not remove the ownership and persistent reservations that Data ONTAP wrote to that LUN. For example, the Hitachi USP storage array does not have a user command for removing persistent reservations from LUNs. If you do not remove persistent reservations through Data ONTAP before removing the Data ONTAP system from service, you must call Hitachi technical support to remove the reservations. Contact technical support for instructions about how to remove persistent reservations from LUNs before removing a Data ONTAP system from service. 80 | Physical Storage Management Guide
Securing data at rest with Storage Encryption You can protect your data at rest from unauthorized access with Storage Encryption. To use this feature, you must understand what Storage Encryption is and how it works, how to set it up and manage it, and how you can destroy data on disks using Storage Encryption. Introduction to Storage Encryption Overview of Storage Encryption concepts, functionality, benefits, and limitations. What Storage Encryption is Storage Encryption is an optional feature that you can enable for additional data protection. It is available on certain supported storage controllers and disk shelves that contain disks with built-in encryption functionality. In a standard storage environment, data is written to disk in cleartext format. This makes the data vulnerable to potential exposure to unauthorized users when disks removed from a storage system are lost or stolen. When you enable Storage Encryption, the storage system protects your data at rest by storing it on self-encrypting disks. The authentication keys used by the self-encrypting disks are stored securely on external key management servers. Purpose of the external key management server An external key management server is a third-party system in your storage environment that securely manages authentication keys used by the self-encrypting disks in the storage system. You link the external key management server to other systems that use authentication or encryption keys such as your storage system. The storage system uses a secure SSL connection to connect to the external key management server to store and retrieve authentication keys. The communication between the storage system and key management server uses the Key Management Interoperability Protocol (KMIP). The external key management server securely stores authentication or encryption keys entrusted to it and provides them upon demand to authorized linked systems. This provides an additional level of security by storing authentication keys separate from the storage system. Additionally, authentication keys are always handled and stored securely. The keys are never displayed in cleartext. You must link at least one key management server to the storage system during the Storage Encryption setup and configuration process. You should link multiple key management servers for redundancy. If the only key management server in the environment becomes unavailable, access to protected data might become unavailable until the key management server is available again. For 81
example, when the storage system needs to unlock self-encrypting disks but cannot retrieve the authentication key from the key management server because it is unavailable. You can specify up to four key servers during or after setup for redundancy. For a list of supported key management servers, see the Interoperability Matrix. How Storage Encryption works Storage Encryption occurs at the firmware level of disks that are equipped with special firmware and hardware to provide the additional security, also known as self-encrypting disks (SEDs). SEDs can operate either in unprotected mode like regular disks, or in protected mode requiring authentication after the power-on process. SEDs always encrypt data for storage. In unprotected mode, the encryption key needed to decrypt and access the data is freely available. In protected mode, the encryption key is protected and requires authentication to be used. When you first enable and configure Storage Encryption on a storage system using SEDs, you create an authentication key that the storage system uses to authenticate itself to the SEDs. You configure the storage system with the IP address to one or more external key management servers that securely stores the authentication key. The storage system communicates with the key management servers at boot time to retrieve the authentication keys. Data ONTAP requires the authentication keys to authenticate itself to the SEDs any time after the SEDs are power-cycled. If the authentication is successful, the SEDs are unlocked. The SEDs use the authentication key to decrypt the data encryption keys stored inside the disk. When presented with a read request, SEDs automatically decrypt the stored data before passing it on to the storage system. When presented with a write request from the storage system, SEDs automatically encrypt the data before writing the data to the disk's storage platters. When the SEDs are locked, Data ONTAP must successfully authenticate itself to the disk before the SEDs allow data to be read or written. When locked, SEDs require authentication each time the disk is powered on. Encryption and decryption happens without a perceptible disk performance decrease or boot time increase. Storage Encryption does not require a separate license key. The only additional required component is an external key management server. When you halt and power down the storage system, including the disk shelves containing SEDs, the disks are locked again and the data becomes inaccessible. Disk operations with SEDs Most of the disk-related operations are identical for SEDs and regular disks. Because storage encryption happens at a very low level, specifically the disk firmware, it does not affect any higher level functionality. The storage controller sees SEDs the same as regular disks, and all functionality remains the same. There are some additional options and requirements with SEDs: 82 | Physical Storage Management Guide
• Sanitizing disks There are additional options to sanitize disks when using SEDs. • Destroying disks An additional option enables you to make the disks permanently inaccessible. Benefits of using Storage Encryption There are several scenarios where using Storage Encryption provides significant benefits by protecting data from unauthorized access when disks removed from a storage system have fallen into the wrong hands. Data protection in case of disk loss or theft Storage Encryption protects your data if disks are lost or stolen. Someone who comes into possession of disks that store data using Storage Encryption cannot access the data. Without the authentication key that is required to authenticate and unlock the disks, all attempts to read or write data result in an error message returned by the SEDs. Circumventing the disk authentication by moving the platters into another disk without encryption firmware would be unsuccessful as well. The data stored on the platters appears as ciphertext and is fully protected from unauthorized access. Data protection when returning disks to vendors Storage Encryption protects your data when you return disks to vendors. The following three options are available to protect data on disks that are removed from a storage system and returned to a vendor: • If the SED is owned by a storage system, it requires authentication to access the data. Since the vendor does not know, or have access to, the authentication key, the vendor cannot access data on the disk. • If you sanitize the disk before returning it to a vendor, it changes the encryption key to a new unknown key. Any subsequent attempts to read data from the disk result in random data. • If you "destroy" the disk, it changes the encryption key to a random unknown key, it changes the authentication key to a random unknown key, and permanently locks the disk, preventing any further decryption of the data and access to the disk. Related tasks Sanitizing disks using Storage Encryption before return to vendor on page 102 Securing data at rest with Storage Encryption | 83
Data protection when moving disks to end-of-life Storage Encryption protects your data when moving a disk to an end-of-life state. You can protect data on a disk by changing the authentication key to a random value that is not stored and permanently locking the drive. This prevents any further decryption of the data and access to the disk. Related tasks Setting the state of disks using Storage Encryption to end-of-life on page 103 Data protection through emergency data shredding Storage Encryption protects your data in emergency situations by allowing you to instantaneously prevent access to the data on the disk. This might include extreme scenarios where power to the storage system or the key management server (or both) is not available, or one or both have fallen into possession of a hostile third-party. Related tasks Emergency shredding of data on disks using Storage Encryption on page 104 Limitations of Storage Encryption You must keep certain limitations in mind when using Storage Encryption. • For the latest information about which storage systems, disk shelves, and key management servers are supported with Storage Encryption, see the Interoperability Matrix. • All disks in the storage system and optional attached disk shelves must have encryption functionality to be able to use Storage Encryption. You cannot mix regular non-encrypting disks with self-encrypting disks. • Storage Encryption key_manager commands are only available for local nodes. They are not available in takeover mode for partner nodes. • Do not configure Storage Encryption to use 10 Gigabit network interfaces for communication with key management servers. This limitation does not apply to serving data. • Storage Encryption supports a maximum of 128 authentication keys per key management server. You receive a warning when the number of stored authentication keys reaches 100. You cannot create new authentication keys when the number of stored authentication keys reaches the limit of 128. You must then delete unused authentication keys before you can create new ones. • Storage Encryption supports KMIP 1.0 and 1.1 for communication with key management servers. 84 | Physical Storage Management Guide
Related information Interoperability Matrix Tool: mysupport.netapp.com/matrix Setting up Storage Encryption During initial setup, your storage system checks whether it is properly configured with self- encrypting disks and is running a version of Data ONTAP that supports Storage Encryption. If the check is successful, you can then launch the Storage Encryption setup wizard after completion of the storage system setup wizard. Information to collect before configuring Storage Encryption You must gather certain information to successfully set up Storage Encryption on your storage system. Information to collect Details Required Optional Network interface name You must provide the name of the network interface the storage system should use to communicate with external key management servers. Note: Do not configure 10-Gigabit network interfaces for communication with key management servers. x Network interface IP address You must provide the IP address of the network interface. This IP address can be in either IPv4 or IPv6 format. x IPv6 network prefix length For IPv6 addresses, you must provide the network prefix length. You can either provide it by appending a slash (/) and the network prefix length directly to the IPv6 address when entering it, or you can enter the network prefix length separately when prompted after entering the address. x Network interface subnet mask You must provide the subnet mask of the network interface. x Network interface gateway IP address You must provide the IP address for the network interface gateway. x Securing data at rest with Storage Encryption | 85
Information to collect Details Required Optional IP addresses for external key management servers You must link the storage system to at least one external key management server during setup. You should add two or more external key management servers to prevent having a single point of failure. If you add only one external key management server and it fails, you can lose access to your data. If you specify IPv6 addresses for external key management servers, you must also provide an IPv6 address for the storage system network interface. x IP address for additional external key management servers You can link the storage system to multiple additional external key management servers during setup for redundancy. x Port number for each external key management server You must provide the port number that each key management server listens on. The port number must be the same for all key management servers. x Public SSL certificate for storage system You must provide a public SSL certificate for the storage system to link it to the external key management server. x Private SSL certificate for storage system You must provide a private SSL certificate for the storage system. x Public SSL certificate for external key management servers You must provide a public SSL certificate for each external key management server to link it to the storage system. x Key tag name You can provide a name that is used to identify all keys belonging to a particular storage system. The default key tag name is the system's host name. x Using SSL for secure key management communication The storage system and key management servers use SSL connections to keep the communication between them secure. This requires you to obtain and install various SSL certificates for the storage system and each key management server before you can set up and configure Storage Encryption. To avoid issues when installing SSL certificates, you should first synchronize the time between the following systems: • The server creating the certificates 86 | Physical Storage Management Guide
• The key management servers • The storage system Requirements for SSL certificates Before obtaining and installing SSL certificates, you must understand what certificates are required and their requirements. SSL certificates for Storage Encryption must use the Privacy Enhanced Mail (PEM) Base-64 encoded X.509 format and follow a strict naming convention. The following table describes the required certificate types and naming conventions: Certificate for... Certificate type Certificate file name Storage system Public client.pem Storage system Private client_private.pem Key management server Public key_management_server_ipaddress_CA.pem key_management_server_ipaddress must be identical to the IP address of the key management server that you use to identify it when running the Storage Encryption setup program. These public and private certificates are required for the storage system and key management servers to establish secure SSL connections with each other and verify each other's identities. The certificates for the storage system are only used by the storage system's KMIP client. The private certificate can be passphrase protected during creation. In this case, the Storage Encryption setup program prompts you to enter the passphrase. If your key management server does not accept self-signed certificates, you also need to include the necessary certificate authority (CA) public certificate. In an HA pair, both nodes must use the same public and private certificates. If you want multiple HA pairs that are connected to the same key management server to have access to each other's keys, all nodes in all HA pairs must use the same public and private certificates. Installing SSL certificates on the storage system You install the necessary SSL certificates on the storage system by using the keymgr install cert command. The SSL certificates are required for securing the communication between the storage system and key management servers. Before you begin You must have obtained the public and private certificates for the storage system and the public certificate for the key management server and named them as required. Securing data at rest with Storage Encryption | 87
Steps 1. Access the nodeshell: system node run -node node_name 2. Copy the certificate files to a temporary location on the storage system. 3. Install the public certificate of the storage system by entering the following command at the storage system prompt: keymgr install cert /path/client.pem 4. Install the private certificate of the storage system by entering the following command at the storage system prompt: keymgr install cert /path/client_private.pem 5. Install the public certificate of the key management server by entering the following command at the storage system prompt: keymgr install cert /path/key_management_server_ipaddress_CA.pem 6. If you are linking multiple key management servers to the storage system, repeat the previous step for each public certificate of each key management server. 7. Exit the nodeshell and return to the clustershell: exit Running the Storage Encryption setup wizard You launch the Storage Encryption setup wizard by using the key_manager setup command. You should run the Storage Encryption setup wizard after you complete setup of the storage system and the storage volumes or when you need to change Storage Encryption settings after initial setup. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. Enter the following command at the storage system prompt: key_manager setup 3. Complete the steps in the wizard to configure Storage Encryption. 4. Exit the nodeshell and return to the clustershell by entering the following command: exit 88 | Physical Storage Management Guide
Example The following command launches the Storage Encryption setup wizard and shows an example of how to configure Storage Encryption: storage-system*> key_manager setup Found client certificate file client.pem. Registration successful for client.pem. Found client private key file client_private.pem. Is this file protected by a passphrase? [no]: Registration successful for client_private.pem. Enter the IP address for a key server, 'q' to quit: 172.22.192.192 Enter the IP address for a key server, 'q' to quit: q Enter the TCP port number for kmip server [6001] : You will now be prompted to enter a key tag name. The key tag name is used to identify all keys belonging to this Data ONTAP system. The default key tag name is based on the system's hostname. Would you like to use <storage-system> as the default key tag name? [yes]: Registering 1 key servers... Found client CA certificate file 172.22.192.192_CA.pem. Registration successful for 172.22.192.192_CA.pem. Registration complete. You will now be prompted for a subset of your network configuration setup. These parameters will define a pre-boot network environment allowing secure connections to the registered key server(s). Enter network interface: e0a Enter IP address: 172.16.132.165 Enter netmask: 255.255.252.0 Enter gateway: 172.16.132.1 Do you wish to enter or generate a passphrase for the system's encrypting drives at this time? [yes]: yes Would you like the system to autogenerate a passphrase? [yes]: yes Key ID: 080CDCB20000000001000000000000003FE505B0C5E3E76061EE48E02A29822C Make sure that you keep a copy of your passphrase, key ID, and key tag name in a secure location in case it is ever needed for recovery purposes. Should the system lock all encrypting drives at this time? yes Securing data at rest with Storage Encryption | 89
Completed rekey on 4 disks: 4 successes, 0 failures, including 0 unknown key and 0 authentication failures. Completed lock on 4 disks: 4 successes, 0 failures, including 0 unknown key and 0 authentication failures. Managing Storage Encryption You can perform various tasks to manage Storage Encryption, including viewing and removing key management servers, and creating, deleting, restoring and synchronizing authentication keys. Adding key management servers You can use the key_manager add command to link key management servers to the storage system. This enables you to add additional key management servers for redundancy after initial setup or to replace existing key management servers. Before you begin You must first install the required storage system and key management server SSL certificates. If they are not present, the command fails. You must know the IP address for each key management server you want to link. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. To add a key management server, enter the following command: key_manager add -key_server key_server_ip_address 3. Exit the nodeshell and return to the clustershell by entering the following command: exit Example The following command adds a link from the storage system to the key management server with the IP address 172.16.132.118: storage-system> key_manager add -key_server 172.16.132.118 Found client certificate file client.pem. Registration successful for client.pem. Found client private key file client_private.pem. Is this file protected by a passphrase? [no]: no Registration successful for client_private.pem. 90 | Physical Storage Management Guide
Registering 1 key servers... Found client CA certificate file 172.16.132.118_CA.pem. Registration successful for 172.16.132.118_CA.pem. Registration complete. Verifying key management server links You use the key_manager status or key_manager query commands to verify that all key management servers are successfully linked to the storage system. These commands are useful for verifying proper operation and troubleshooting. About this task Both commands display whether key management servers are responding. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. Perform one of the following actions: If you want to... Then enter the following command: Check the status of a specific key management server key_manager status - key_server key_server_ip_address Check the status of all key management servers key_manager status Check the status of all key management servers and view additional server details. key_manager query The key_manager query command displays additional information about key tags and key IDs. 3. Check the output to verify that all of the appropriate keys are available in the Data ONTAP key table. If the output of the key_manager query command displays key IDs marked with an asterisk (*), those keys exist on a key server but are not currently available in the Data ONTAP key table. To import those keys from the key management server into the key table, enter the following command: key_manager restore 4. Exit the nodeshell and return to the clustershell by entering the following command: exit Securing data at rest with Storage Encryption | 91
Examples The following command checks the status of all key management servers linked to the storage system: storage-system> key_manager status Key server Status 172.16.132.118 Server is responding 172.16.132.211 Server is responding The following command checks the status of all key management servers linked to the storage system and displays additional information: storage-system> key_manager query Key server 172.16.132.118 is responding. Key server 172.16.132.211 is responding. Key server 172.16.132.118 reports 4 keys. Key tag Key ID -------- ------- storage-system 080CDCB20... storage-system 080CDCB20... storage-system 080CDCB20... storage-system 080CDCB20... Key server 172.16.132.211 reports 4 keys. Key tag Key ID -------- ------- storage-system *080CDCB20... storage-system 080CDCB20... storage-system 080CDCB20... storage-system *080CDCB20... Displaying key management server information You can display information about the external key management servers associated with the storage system by using the key_manager show command. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. To display external key management servers, enter the following command: key_manager show 92 | Physical Storage Management Guide
All external key management servers associated with the storage system are listed. 3. Exit the nodeshell and return to the clustershell by entering the following command: exit Example The following command displays all external key management servers associated with the storage system: storage-system> key_manager show 172.18.99.175 Removing key management servers If you no longer want to use a key management server to store authentication keys used by self- encrypting disks in the storage system, you can remove the key management server link to the storage system by using the key_manager remove command. Before you begin You must know the IP address for each key management server that you want to remove. About this task Storage Encryption requires at least one key management server linked to the storage system to operate. If you want to replace a single key management server with another one, you must first add the new one before removing the old one. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. To remove key management servers, enter the following command: key_manager remove -key_server key_server_ip_address -key_server key_server_ip_address specifies the IP address of the key management server you want to remove. 3. Exit the nodeshell and return to the clustershell by entering the following command: exit Securing data at rest with Storage Encryption | 93
Example The following command removes the link between the storage system and the key management server with the IP address 172.18.99.175: storage-system> key_manager remove -key_server 172.18.99.175 Key server 172.18.99.175 will be unregistered from service. Unregistration successful. What happens when key management servers are not reachable during the boot process Data ONTAP takes certain precautions to avoid undesired behavior in the event that the storage system cannot reach any of the specified key management servers during the boot process. If the storage system is configured for Storage Encryption, the SEDs are rekeyed and locked, and the SEDs are powered on, the storage system must retrieve the required authentication keys from the key management servers to authenticate itself to the SEDs before it can access the data. The storage system attempts to contact the specified key management servers for up to three hours. If the storage system cannot reach any of them after that time, the boot process stops and the storage system halts. If the storage system successfully contacts any specified key management server, it then attempts to establish an SSL connection for up to 15 minutes. If the storage system cannot establish an SSL connection with any specified key management server, the boot process stops and the storage system halts. While the storage system attempts to contact and connect to key management servers, it displays detailed information about the failed contact attempts at the CLI. You can interrupt the contact attempts at any time by pressing Ctrl-C. As a security measure, SEDs allow only a limited number of unauthorized access attempts, after which they disable access to the existing data. If the storage system cannot contact any specified key management servers to obtain the proper authentication keys, it can only attempt to authenticate with the default key which leads to a failed attempt and a panic. If the storage system is configured to automatically reboot in case of a panic, it enters a boot loop which results in continuous failed authentication attempts on the SEDs. Halting the storage system in these scenarios is by design to prevent the storage system from entering a boot loop and possible unintended data loss as a result of the SEDs locked permanently due to exceeding the safety limit of a certain number of consecutive failed authentication attempts. The limit and the type of lockout protection depends on the type of SED: 94 | Physical Storage Management Guide
SED type Number of consecutive failed authentication attempts resulting in lockout Lockout protection type when safety limit is reached HDD 1024 Permanent. Data cannot be recovered, even when the proper authentication key becomes available again. SSD 5 Temporary. Lockout is only in effect until disk is power-cycled. For all SED types, a successful authentication resets the try count to zero. If you encounter this scenario where the storage system is halted due to failure to reach any specified key management servers, you must first identify and correct the cause for the communication failure before you attempt to continue booting the storage system. Displaying Storage Encryption disk information You can display information about self-encrypting disks by using the disk encrypt show command. This command displays the key ID and lock status for each self-encrypting disk. About this task The key ID displayed in the command output is an identifier used by Storage Encryption and key management servers as a reference to the authentication key. It is not the actual authentication key or the data encryption key. Steps 1. Access the nodeshell: system node run -node node_name 2. To display information about SEDs: disk encrypt show The disk encrypt show, lock, and rekey commands support extended wildcard matching. For more information, see the disk encrypt show man page. 3. Exit the nodeshell and return to the clustershell: exit Example The following command displays the status of each self-encrypting disk: Securing data at rest with Storage Encryption | 95
storage-system> disk encrypt show Disk Key ID Locked? 0c.00.1 080CF0C8000000000100000000000000A948EE8604F4598ADFFB185B5BB7FED3 Yes 0c.00.0 080CF0C8000000000100000000000000A948EE8604F4598ADFFB185B5BB7FED3 Yes 0c.00.3 080CF0C8000000000100000000000000A948EE8604F4598ADFFB185B5BB7FED3 Yes 0c.00.4 080CF0C8000000000100000000000000A948EE8604F4598ADFFB185B5BB7FED3 Yes 0c.00.2 080CF0C8000000000100000000000000A948EE8604F4598ADFFB185B5BB7FED3 Yes 0c.00.5 080CF0C8000000000100000000000000A948EE8604F4598ADFFB185B5BB7FED3 Yes Changing the authentication key You can change the authentication key at any time by using the key_manager rekey command. You might want to change the authentication key as part of your security protocol or when moving an aggregate to another storage system. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. Perform one of the following actions: If you want to... Then... Change the authentication key and enter a new one manually a. Enter the following command at the storage system prompt: key_manager rekey -manual -key_tag key_tag b. When prompted, enter the new authentication key. It must be 20 to 32 characters long. Change the authentication key and have the system generate a new one automatically Enter the following command at the storage system prompt: key_manager rekey -key_tag key_tag key_tag is the label used to associate keys with a particular storage system. If you do not specify a key tag, the storage system uses the key tag specified when you set up Storage Encryption. If you did not specify this key tag during setup, it uses the parent key tag as the default. Each node has a parent key tag. HA pair members share the same parent key tag. 3. Exit the nodeshell and return to the clustershell by entering the following command: exit Example The following command changes the authentication key and prompts you to enter a new one manually. You can run the disk encrypt show command after completion to verify the results. 96 | Physical Storage Management Guide
storage-system> key_manager rekey -manual Please enter a new passphrase: Please reenter the new passphrase: New passphrase generated. Key ID: 080CDCB2000000000100000000000000B0A11CBF3DDD20EFB0FBB5EE198DB22A Key tag: storage-system Notice: Remember to store the passphrase and the Key ID in a secure location. Passphrase, key ID, and key tag synchronized with the following key server(s): 172.16.132.118 172.16.132.211 Completed rekey on 4 disks: 4 successes, 0 failures, including 0 unknown key and 0 authentication failures. Retrieving authentication keys You can use the key_manager restore command to retrieve authentication keys from a key management server to a storage system. For example, when you created authentication keys on a node, you use this command to retrieve the keys for use on the partner node. Before you begin You must know the IP address for each key management server that you want to retrieve authentication keys from. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. To retrieve authentication keys from a key management server to the storage system, enter the following command: key_manager restore -key_server key_server_ip_address -key_tag key_tag If all specified key management servers are available, you can use the -all option instead of the -key_server option to clear out the current Data ONTAP key table and retrieve all keys matching the specified key tag from all specified key management servers. 3. Exit the nodeshell and return to the clustershell by entering the following command: exit Securing data at rest with Storage Encryption | 97
Examples The following command restores keys with the key tag storage-system from the key management server with the IP address 172.18.99.175: storage-system> key_manager restore -key_server 172.18.99.175 - key_tag storage-system The following command restores all keys with the key tag storage-system from all key management servers linked to the storage system: storage-system> key_manager restore -all -key_tag storage-system Deleting an authentication key You can delete an authentication key that is no longer needed by removing it from the external key management server. Before you begin Verify that the authentication key is no longer needed before deleting it. Deleting an authentication key that is still in use can permanently prevent access to data on a storage system. Step 1. Refer to the documentation for the external key management server for details on how to delete stored authentication keys. SSL issues due to expired certificates If the SSL certificates used to secure key management communication between the storage system and key management servers expire, the storage system can no longer retrieve authentication keys from the key management server at bootup. This issue can cause data on SEDs to be unavailable. You can prevent this issue by updating all SSL certificates before their individual expiration dates. SSL certificates have a limited lifespan because they have an expiration date. After the SSL certificates reach their expiration dates, the certificates are no longer valid. When this happens, SSL connections that use expired certificates fail. For Storage Encryption, this means that the SSL connections between the storage system and the key management servers fail, the storage system no longer can retrieve authentication keys when needed, and data access to the SEDs fails, resulting in storage system panic and downtime. To prevent this issue from occurring, you must keep track of the expiration dates of all installed SSL certificates so that you can obtain new SSL certificates before they expire. 98 | Physical Storage Management Guide
After you have obtained the new certificate files, you must first remove the existing certificate files from the storage system, and then install the new certificates on the storage system. Steps 1. Removing old SSL certificates before installing new ones on page 99 If you want to update or reinstall the SSL certificates used by Storage Encryption, you must first manually remove the old ones to ensure that the new ones are used. 2. Installing replacement SSL certificates on the storage system on page 99 After you remove the old certificates, you create the new replacement SSL certificates, save them with the proper file name and format, and then install them on the storage system. Removing old SSL certificates before installing new ones If you want to update or reinstall the SSL certificates used by Storage Encryption, you must first manually remove the old ones to ensure that the new ones are used. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. Remove the IP addresses of all key management servers by entering the following command for each key management server: key_manager remove -key_server key_server_ip_address 3. Remove the storage system's client certificates by entering the following commands: keymgr delete cert client_private.pem keymgr delete cert client.pem 4. Remove all installed key management server certificates by entering the following commands for each key management server: keymgr delete cert key_server_ip_address_CA.pem 5. Exit the nodeshell and return to the clustershell by entering the following command: exit Installing replacement SSL certificates on the storage system After you remove the old certificates, you create the new replacement SSL certificates, save them with the proper file name and format, and then install them on the storage system. Before you begin • You must have removed the old certificates that are about to expire from the storage system. Securing data at rest with Storage Encryption | 99
• You must have obtained the replacement public and private certificates for the storage system and the public certificate for the key management server, and named them as required. For more information, see the Clustered Data ONTAP System Administration Guide for Cluster Administrators. • You must have installed the appropriate new certificates on the key management server. For more information, see the documentation for your key management server. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. Copy the certificate files to a temporary location on the storage system. 3. Install the public certificate of the storage system by entering the following command: keymgr install cert /path/client.pem 4. Install the private certificate of the storage system by entering the following command: keymgr install cert /path/client_private.pem 5. Install the public certificate of all key management servers by entering the following command for each key management server: keymgr install cert /path/key_management_server_ipaddress_CA.pem 6. Add all key management servers by entering the following command for each key management server: key_manager add -key_server key_server_ip_address 7. Verify connectivity between the storage system and key management servers by entering the following command: key_manager query You should see a list of existing key IDs retrieved from the key management servers. 8. Exit the nodeshell and return to the clustershell by entering the following command: exit Returning SEDs to unprotected mode If your storage system is configured to use Storage Encryption but you decide to stop using this feature, you can do so by returning the SEDs to unprotected mode. You cannot disable Storage Encryption altogether because SEDs always encrypt data for storage. However, you can return them 100 | Physical Storage Management Guide
to unprotected mode where they no longer use secret authentication keys, and use the default MSID instead. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. To change the authentication key for all SEDs on the storage system back to the default MSID, enter the following command: disk encrypt rekey * 0x0 3. If you expect to operate the storage system in unprotected mode permanently, you should also remove all key management servers by entering the following command for each one: key_manager remove -key_server key_server_ip_address -key_server key_server_ip_address specifies the IP address of the key management server you want to remove. The storage system displays two kmip_init errors during every bootup after you remove all key management servers. These errors are normal in this situation and you can disregard them. 4. If you expect to operate the storage system in unprotected mode permanently and you removed all key management servers in the preceding step, you should view the list of installed Storage Encryption related SSL certificates, and then remove all key management server SSL certificates: keymgr cert list keymgr delete cert client.pem keymgr delete cert client_private.pem keymgr delete cert key_management_server_ipaddress_CA.pem If you had multiple key management servers linked to the storage system, repeat the last command for each public certificate of each key management server. 5. Exit the nodeshell and return to the clustershell by entering the following command: exit Securing data at rest with Storage Encryption | 101
Destroying data on disks using Storage Encryption You can destroy data stored on disks using Storage Encryption for security reasons, including sanitizing the disks, setting the disk state to end-of-life, and emergency shredding of the data. Sanitizing disks using Storage Encryption before return to vendor If you want to return a disk to a vendor but do not want anyone to access sensitive data on the disk, you can sanitize it first by using the disk encrypt sanitize command. This renders the data on the disk inaccessible, but the disk can be reused. This command only works on spare disks. Steps 1. Migrate any data that needs to be preserved to a different aggregate. 2. Destroy the aggregate. 3. Access the nodeshell by entering the following command: system node run -node node_name 4. Identify the disk ID for the disk to be sanitized by entering the following command: disk encrypt show 5. Enter the following command: disk encrypt sanitize disk_ID 6. Exit the nodeshell and return to the clustershell by entering the following command: exit Example The following command sanitizes a self-encrypting disk with the disk ID 0c.00.3. You can run the sysconfig -r command before and after the operation to verify the results. storage-system> sysconfig -r Aggregate aggr0 (online, raid_dp) (block checksums) Plex /aggr0/plex0 (online, normal, active) RAID group /aggr0/plex0/rg0 (normal) RAID Disk Device HA SHELF BAY CHAN Pool Type RPM Used (MB/blks) Phys (MB/blks) --------- ------ ------------- ---- ---- ---- ----- -------------- -------------- dparity 0c.00.0 0c 0 0 SA:B - SAS 15000 560000/1146880000 560208/1147307688 parity 0c.00.1 0c 0 1 SA:B - SAS 15000 560000/1146880000 560208/1147307688 data 0c.00.2 0c 0 2 SA:B - SAS 15000 560000/1146880000 560208/1147307688 Spare disks RAID Disk Device HA SHELF BAY CHAN Pool Type RPM Used (MB/blks) Phys (MB/blks) --------- ------ ------------- ---- ---- ---- ----- -------------- -------------- Spare disks for block or zoned checksum traditional volumes or aggregates spare 0c.00.3 0c 0 3 SA:B - SAS 15000 560000/1146880000 560208/1147307688 spare 0c.00.4 0c 0 4 SA:B - SAS 15000 560000/1146880000 560208/1147307688 102 | Physical Storage Management Guide
spare 0c.00.5 0c 0 5 SA:B - SAS 15000 560000/1146880000 560208/1147307688 storage-system> disk encrypt sanitize 0c.00.3 storage-system> Wed Jun 30 17:49:16 PDT [disk.failmsg:error]: Disk 0c.00.3 (3SL04F3V00009015WTHU): message received. Wed Jun 30 17:49:16 PDT [raid.disk.unload.done:info]: Unload of Disk 0c.00.3 Shelf 0 Bay 3 [SYSTEM X415_S15K7560A15 NQS3] S/N [3SL04F3V00009015WTHU] has completed successfully storage-system> Wed Jun 30 17:49:25 PDT [disk.sanit.complete:info]: Disk 0c.00.3 [S/N 3SL04F3V00009015WTHU] has completed sanitization. storage-system> sysconfig - r Aggregate aggr0 (online, raid_dp) (block checksums) Plex /aggr0/plex0 (online, normal, active) RAID group /aggr0/plex0/rg0 (normal) RAID Disk Device HA SHELF BAY CHAN Pool Type RPM Used (MB/blks) Phys (MB/blks) --------- ------ ------------- ---- ---- ---- ----- -------------- -------------- dparity 0c.00.0 0c 0 0 SA:B - SAS 15000 560000/1146880000 560208/1147307688 parity 0c.00.1 0c 0 1 SA:B - SAS 15000 560000/1146880000 560208/1147307688 data 0c.00.2 0c 0 2 SA:B - SAS 15000 560000/1146880000 560208/1147307688 Spare disks RAID Disk Device HA SHELF BAY CHAN Pool Type RPM Used (MB/blks) Phys (MB/blks) --------- ------ ------------- ---- ---- ---- ----- -------------- -------------- Spare disks for block or zoned checksum traditional volumes or aggregates spare 0c.00.4 0c 0 4 SA:B - SAS 15000 560000/1146880000 560208/1147307688 spare 0c.00.5 0c 0 5 SA:B - SAS 15000 560000/1146880000 560208/1147307688 Maintenance disks RAID Disk Device HA SHELF BAY CHAN Pool Type RPM Used (MB/blks) Phys (MB/blks) --------- ------ ------------- ---- ---- ---- ----- -------------- -------------- sanitized 0c.00.3 0c 0 3 SA:B - SAS 15000 560000/1146880000 560208/1147307688 storage-system> Setting the state of disks using Storage Encryption to end-of-life If you want to render a disk permanently unusable and the data on it inaccessible, you can set the state of the disk to end-of-life by using the disk encrypt destroy command. This command only works on spare disks. Steps 1. Remove any data from the aggregate containing the disk. 2. Migrate any data that needs to be preserved to a different aggregate. 3. Destroy the aggregate. 4. Access the nodeshell: system node run -node node_name 5. Enter the following command: disk encrypt destroy disk_ID 6. Exit the nodeshell and return to the clustershell: exit Securing data at rest with Storage Encryption | 103
7. If the disk has a PSID printed on its label but you do not want the disk to be able to be reset to factory settings and returned to service at a later time, obliterate all instances of the PSID on the disk label (text or scannable code). Be sure to also destroy any copies, scans, or photographs of the disk label. Result The disk's encryption key is set to an unknown random value and the disk is irreversibly locked. The disk is now completely unusable and can be safely disposed of without risk of unauthorized data access. Emergency shredding of data on disks using Storage Encryption In case of a security emergency, you can instantly prevent access to data on disks using Storage Encryption, even if power is not available to the storage system or the external key server. Before you begin You must configure the external key server so that it only operates if an easily destroyed authentication item (for example, a smart card or USB drive) is present. See the documentation for the external key management server for details. About this task The steps for emergency shredding vary depending on whether power is available to the storage system and the external key server. Step 1. Perform one of the following actions: If... Then... Power is available to the storage system and you have time to gracefully take the storage system offline a. If the storage system is a node in an HA pair, disable takeover. b. Take all aggregates offline and destroy them. c. Halt the storage system. d. Boot into maintenance mode. e. Enter the following command: disk encrypt sanitize -all This leaves the storage system in a permanently disabled state with all data erased. To use the storage system again, you must set it up from the beginning. 104 | Physical Storage Management Guide
If... Then... Power is available to the storage system and you must shred the data immediately; time is critical a. If the storage system is a node in an HA pair, disable takeover. b. Access the nodeshell by entering the following command: system node run -node node_name c. Set the privilege level to advanced. d. Enter the following command: disk encrypt sanitize -all The storage system panics, which is expected due to the abrupt nature of the procedure. It leaves the storage system in a permanently disabled state with all data erased. To use the storage system again, you must set it up from the beginning. Power is available to the external key server but not to the storage system a. Log in to the external key server. b. Destroy all keys associated with the disks containing data to protect. Power is not available to the external key server or the storage system Destroy the authentication item for the key server (for example, the smart card). If power to the systems is restored, the external key server cannot operate due to the missing authentication item. This prevents access to the disk encryption keys by the storage system, and therefore access to the data on the disks. What function the physical secure ID has for SEDs Certain self-encrypting disks (SEDs) feature additional functionality to reset the disk to factory settings. These disks have a physical secure ID (PSID) printed on the disk label that is required to perform a factory reset. The PSID is unique to each drive. It is printed on the disk label and visible to anyone with physical access to the SED. The PSID is not electronically retrievable from the SED. If the disk label is obliterated, the PSID is lost and cannot be recovered. Using the PSID to perform a factory reset causes all disk parameters to be reset to factory original settings, including the following: • The encryption key used to encrypt and decrypt the data on the media is changed to an unknown value. If the SED contained data, access to the data is permanently lost. The new unknown encryption key cannot be retrieved. This operation cannot be undone. • The authentication key required to authenticate to the SED is changed back to the manufacturer's default secure ID (MSID). New data stored on the SED is not protected until the MSID is changed to a new secret authentication key. Securing data at rest with Storage Encryption | 105
• The SED can be returned into service, even if its state was previously set to end-of-life. If the SED was previously used but then set to end-of-life state using the disk encrypt destroy command, use of the PSID recovers the SED from this state and returns it to normal service state. However, it only returns it to factory original settings. It cannot in any way recover previously used encryption or authentication keys or restore access to previous data stored on the SED. SEDs that have PSID functionality There are several models of SEDs but only some of them have PSID functionality. Earlier SED models with PSID functionality do not have the PSID printed on the disk label and therefore cannot use it. Use the following table for guidance to find out whether your SEDs have PSID functionality: Disk model Has PSID functio nality Has PSID printed on disk label Comments X414 No No SED cannot be reset to factory original settings using a PSID. X415 No No SED cannot be reset to factory original settings using a PSID. X416 Yes Maybe Check the label on the physical disk. If a PSID is printed on the label, the SED can be reset to factory original settings using the PSID. X417 Yes Maybe Check the label on the physical disk. If a PSID is printed on the label, the SED can be reset to factory original settings using the PSID. All other SEDs Yes Yes Check the label on the physical disk to obtain the PSID. Resetting an SED to factory original settings If you previously set the state of an SED to end-of-life by using the disk encrypt destroy command but now want to return it to service, you can reset it to its factory original settings by using the disk encrypt revert_original command provided that the disk has a PSID printed on its label. Before you begin You must have obtained the SED's PSID from its disk label. 106 | Physical Storage Management Guide
Steps 1. Access the nodeshell: system node run -node node_name 2. Set the privilege level to advanced: priv set advanced 3. Reset the disk to its factory configured settings: disk encrypt revert_original disk_ID PSID The PSID is the physical secure ID printed on the disk label that is required to perform a factory reset. 4. Make the disk unavailable to the storage system: disk fail disk_name 5. Make the disk available again to the storage system: disk unfail -s disk_name 6. Verify that the operation was successful: sysconfig -r The disk that you reset shows up as a spare disk. 7. Return to the admin privilege level: priv set admin 8. Exit the nodeshell and return to the clustershell: exit Securing data at rest with Storage Encryption | 107
How Data ONTAP uses RAID to protect your data and data availability Understanding how RAID protects your data and data availability can help you administer your storage systems more effectively. For native storage, Data ONTAP uses RAID-DP (double-parity) or RAID Level 4 (RAID4) protection to ensure data integrity within a RAID group even if one or two of those drives fail. Parity drives provide redundancy for the data stored in the data drives. If a drive fails (or, for RAID-DP, up to two drives), the RAID subsystem can use the parity drives to reconstruct the data in the drive that failed. For array LUNs, Data ONTAP stripes data across the array LUNs using RAID0. The storage arrays, not Data ONTAP, provide the RAID protection for the array LUNs that they make available to Data ONTAP. RAID protection levels for disks Data ONTAP supports two levels of RAID protection for aggregates composed of disks in native disk shelves: RAID-DP and RAID4. RAID-DP is the default RAID level for new aggregates. For more information about configuring RAID, see Technical Report 3437: Storage Subsystem Resiliency Guide. Related information TR 3437: Storage Subsystem Resiliency Guide What RAID-DP protection is If an aggregate is configured for RAID-DP protection, Data ONTAP reconstructs the data from one or two failed disks within a RAID group and transfers that reconstructed data to one or two spare disks as necessary. RAID-DP provides double-parity disk protection when the following conditions occur: • There is a single-disk failure or double-disk failure within a RAID group. • There are media errors on a block when Data ONTAP is attempting to reconstruct a failed disk. The minimum number of disks in a RAID-DP group is three: at least one data disk, one regular parity disk, and one double-parity (dParity) disk. However, for non-root aggregates with only one RAID group, you must have at least 5 disks (three data disks and two parity disks). If there is a data-disk failure or parity-disk failure in a RAID-DP group, Data ONTAP replaces the failed disk in the RAID group with a spare disk and uses the parity data to reconstruct the data of the 108 | Physical Storage Management Guide
failed disk on the replacement disk. If there is a double-disk failure, Data ONTAP replaces the failed disks in the RAID group with two spare disks and uses the double-parity data to reconstruct the data of the failed disks on the replacement disks. RAID-DP is the default RAID type for all aggregates. What RAID4 protection is RAID4 provides single-parity disk protection against single-disk failure within a RAID group. If an aggregate is configured for RAID4 protection, Data ONTAP reconstructs the data from a single failed disk within a RAID group and transfers that reconstructed data to a spare disk. The minimum number of disks in a RAID4 group is two: at least one data disk and one parity disk. However, for non-root aggregates with only one RAID group, you must have at least 3 disks (two data disks and one parity disk). If there is a single data or parity disk failure in a RAID4 group, Data ONTAP replaces the failed disk in the RAID group with a spare disk and uses the parity data to reconstruct the failed disk’s data on the replacement disk. If no spare disks are available, Data ONTAP goes into degraded mode and alerts you of this condition. Attention: With RAID4, if there is a second disk failure before data can be reconstructed from the data on the first failed disk, there will be data loss. To avoid data loss when two disks fail, you can select RAID-DP. This provides two parity disks to protect you from data loss when two disk failures occur in the same RAID group before the first failed disk can be reconstructed. RAID protection for array LUNs Storage arrays provide the RAID protection for the array LUNs that they make available to Data ONTAP; Data ONTAP does not provide the RAID protection. Data ONTAP uses RAID0 (striping) for array LUNs. Data ONTAP supports a variety of RAID types on the storage arrays, except RAID0 because RAID0 does not provide storage protection. When creating RAID groups on storage arrays, you need to follow the best practices of the storage array vendor to ensure that there is an adequate level of protection on the storage array so that disk failure does not result in loss of data or loss of access to data. Note: A RAID group on a storage array is the arrangement of disks that together form the defined RAID level. Each RAID group supports only one RAID type. The number of disks that you select for a RAID group determines the RAID type that a particular RAID group supports. Different storage array vendors use different terms to describe this entity—RAID groups, parity groups, disk groups, Parity RAID groups, and other terms. Note: Data ONTAP supports RAID4 and RAID-DP on native disk shelves but supports only RAID0 on array LUNs. How Data ONTAP uses RAID to protect your data and data availability | 109
RAID protection for Data ONTAP-v storage Because Data ONTAP-v storage is connected to the host server, rather than a storage system running Data ONTAP, the host server provides the RAID protection for the physical disks. Data ONTAP uses RAID0 for the virtual disks to optimize performance. See the Data ONTAP Edge Installation and Administration Guide for more information. Protection provided by RAID and SyncMirror Combining RAID and SyncMirror provides protection against more types of drive failures than using RAID alone. You can use RAID in combination with the SyncMirror functionality, which also offers protection against data loss due to drive or other hardware component failure. SyncMirror protects against data loss by maintaining two copies of the data contained in the aggregate, one in each plex. Any data loss due to drive failure in one plex is repaired by the undamaged data in the other plex. For more information about SyncMirror, see the Clustered Data ONTAP Data Protection Guide. The following tables show the differences between using RAID alone and using RAID with SyncMirror: Table 1: RAID-DP and SyncMirror Criteria RAID-DP alone RAID-DP with SyncMirror Failures protected against • Single-drive failure • Double-drive failure within a single RAID group • Multiple-drive failures, as long as no more than two drives within a single RAID group fail • All failures protected against by RAID-DP alone • Any combination of failures protected against by RAID-DP alone in one plex, concurrent with an unlimited number of failures in the other plex • Storage subsystem failures (HBA, cables, shelf), as long as only one plex is affected 110 | Physical Storage Management Guide
Criteria RAID-DP alone RAID-DP with SyncMirror Failures not protected against • Three or more concurrent drive failures within a single RAID group • Storage subsystem failures (HBA, cables, shelf) that lead to three or more concurrent drive failures within a single RAID group • Three or more concurrent drive failures in a single RAID group on both plexes Required resources per RAID group n data drives + 2 parity disks 2 x (n data drives + 2 parity drives) Performance cost Almost none Low mirroring overhead; can improve performance Additional cost and complexity None SyncMirror license and configuration Table 2: RAID4 and SyncMirror Criteria RAID4 alone RAID4 with SyncMirror Failures protected against • Single-disk failure • Multiple-disk failures, as long as no more than one disk within a single RAID group fails • All failures protected against by RAID4 alone • Any combination of failures protected against by RAID4 alone in one plex, concurrent with an unlimited number of failures in the other plex • Storage subsystem failures (HBA, cables, shelf), as long as only one plex is affected How Data ONTAP uses RAID to protect your data and data availability | 111
Criteria RAID4 alone RAID4 with SyncMirror Failures not protected against • Two or more concurrent drive failures within a single RAID group • Storage subsystem failures (HBA, cables, shelf) that lead to two or more concurrent drive failures within a single RAID group • Two or more concurrent drive failures in a single RAID group on both plexes Required resources per RAID group n data drives + 1 parity drive 2 x (n data drives + 1 parity drive) Performance cost None Low mirroring overhead; can improve performance Additional cost and complexity None SyncMirror configuration and extra storage requirement Table 3: RAID0 and SyncMirror Criteria RAID0 alone RAID0 with SyncMirror Failures protected against No protection against any failures RAID protection is provided by the RAID implemented on the storage array. Any combination of array LUN, connectivity, or hardware failures, as long as only one plex is affected Failures not protected against No protection against any failures RAID protection is provided by the RAID implemented on the storage array. Any concurrent failures that affect both plexes Required array LUN resources per RAID group No extra array LUNs required other than n data array LUNs 2 x n data array LUNs Performance cost None Low mirroring overhead; can improve performance Additional cost and complexity None SyncMirror configuration and extra storage requirement 112 | Physical Storage Management Guide
Understanding RAID drive types Data ONTAP classifies drives (or, for partitioned drives, partitions) as one of four types for RAID: data, hot spare, parity, or dParity. You manage disks differently depending on whether they are spare or being used in an aggregate. The RAID type is determined by how RAID is using a drive or partition; it is different from the Data ONTAP disk type. You cannot affect the RAID type for a drive. The RAID type is displayed in the Position column for many storage commands. For drives using root-data partitioning and SSDs in storage pools, a single drive might be used in multiple ways for RAID. For example, the root partition of a partitioned drive might be a spare partition, whereas the data partition might be being used for parity. For this reason, the RAID drive type for partitioned drives and SSDs in storage pools is displayed simply as shared. Data disk Holds data stored on behalf of clients within RAID groups (and any data generated about the state of the storage system as a result of a malfunction). Spare disk Does not hold usable data, but is available to be added to a RAID group in an aggregate. Any functioning disk that is not assigned to an aggregate but is assigned to a system functions as a hot spare disk. Parity disk Stores row parity information that is used for data reconstruction when a single disk drive fails within the RAID group. dParity disk Stores diagonal parity information that is used for data reconstruction when two disk drives fail within the RAID group, if RAID-DP is enabled. How RAID groups work A RAID group consists of one or more data disks or array LUNs, across which client data is striped and stored, and up to two parity disks, depending on the RAID level of the aggregate that contains the RAID group. RAID-DP uses two parity disks to ensure data recoverability even if two disks within the RAID group fail. RAID4 uses one parity disk to ensure data recoverability if one disk within the RAID group fails. How Data ONTAP uses RAID to protect your data and data availability | 113
RAID0 does not use any parity disks; it does not provide data recoverability if any disks within the RAID group fail. How RAID groups are named Within each aggregate, RAID groups are named rg0, rg1, rg2, and so on in order of their creation. You cannot specify the names of RAID groups. Considerations for sizing RAID groups Configuring an optimum RAID group size requires a trade-off of factors. You must decide which factors—speed of RAID rebuild, assurance against risk of data loss due to drive failure, optimizing I/O performance, and maximizing data storage space—are most important for the aggregate that you are configuring. When you create larger RAID groups, you maximize the space available for data storage for the same amount of storage used for parity (also known as the “parity tax”). On the other hand, when a disk fails in a larger RAID group, reconstruction time is increased, impacting performance for a longer period of time. In addition, having more disks in a RAID group increases the probability of a multiple disk failure within the same RAID group. HDD or array LUN RAID groups You should follow these guidelines when sizing your RAID groups composed of HDDs or array LUNs: • All RAID groups in an aggregate should have a similar number of disks. The RAID groups do not have to be exactly the same size, but you should avoid having any RAID group that is less than one half the size of other RAID groups in the same aggregate when possible. • The recommended range of RAID group size is between 12 and 20. The reliability of performance disks can support a RAID group size of up to 28, if needed. • If you can satisfy the first two guidelines with multiple RAID group sizes, you should choose the larger size. SSD RAID groups in Flash Pool aggregates The SSD RAID group size can be different from the RAID group size for the HDD RAID groups in a Flash Pool aggregate. Usually, you should ensure that you have only one SSD RAID group for a Flash Pool aggregate, to minimize the number of SSDs required for parity. SSD RAID groups in SSD aggregates You should follow these guidelines when sizing your RAID groups composed of SSDs: • All RAID groups in an aggregate should have a similar number of drives. 114 | Physical Storage Management Guide
The RAID groups do not have to be exactly the same size, but you should avoid having any RAID group that is less than one half the size of other RAID groups in the same aggregate when possible. • For RAID-DP, the recommended range of RAID group size is between 20 and 28. Related references Storage limits on page 195 Customizing the size of your RAID groups You can customize the size of your RAID groups to ensure that your RAID group sizes are appropriate for the amount of storage you plan to include for an aggregate. About this task For standard aggregates, you change the size of RAID groups on a per-aggregate basis. For Flash Pool aggregates, you can change the RAID group size for the SSD RAID groups and the HDD RAID groups independently. The following list outlines some facts about changing the RAID group size: • By default, if the number of disks or array LUNs in the most recently created RAID group is less than the new RAID group size, disks or array LUNs will be added to the most recently created RAID group until it reaches the new size. • All other existing RAID groups in that aggregate remain the same size, unless you explicitly add disks to them. • You can never cause a RAID group to become larger than the current maximum RAID group size for the aggregate. • You cannot decrease the size of already created RAID groups. • The new size applies to all RAID groups in that aggregate (or, in the case of a Flash Pool aggregate, all RAID groups for the affected RAID group type—SSD or HDD). Step 1. Use the applicable command: If you want to... Enter the following command... Change the maximum RAID group size for the SSD RAID groups of a Flash Pool aggregate storage aggregate modify -aggregate aggr_name - cache-raid-group-size size How Data ONTAP uses RAID to protect your data and data availability | 115
If you want to... Enter the following command... Change the maximum size of any other RAID groups storage aggregate modify -aggregate aggr_name - maxraidsize size Examples The following command changes the maximum RAID group size of the aggregate n1_a4 to 20 disks or array LUNs: storage aggregate modify -aggregate n1_a4 -maxraidsize 20 The following command changes the maximum RAID group size of the SSD cache RAID groups of the Flash Pool aggregate n1_cache_a2 to 24: storage aggregate modify -aggregate n1_cache_a2 -cache-raid-group-size 24 Related concepts Considerations for sizing RAID groups on page 114 Considerations for Data ONTAP RAID groups for array LUNs Setting up Data ONTAP RAID groups for array LUNs requires planning and coordination with the storage array administrator so that the administrator makes the number and size of array LUNs you need available to Data ONTAP. For array LUNs, Data ONTAP uses RAID0 RAID groups to determine where to allocate data to the LUNs on the storage array. The RAID0 RAID groups are not used for RAID data protection. The storage arrays provide the RAID data protection. Note: Data ONTAP RAID groups are similar in concept to what storage array vendors call RAID groups, parity groups, disk groups, Parity RAID groups, and other terms. Follow these steps when planning your Data ONTAP RAID groups for array LUNs: 1. Plan the size of the aggregate that best meets your data needs. 2. Plan the number and size of RAID groups that you need for the size of the aggregate. Note: It is best to use the default RAID group size for array LUNs. The default RAID group size is adequate for most organizations. The default RAID group size is different for array LUNs and disks. 3. Plan the size of the LUNs that you need in your RAID groups. • To avoid a performance penalty, all array LUNs in a particular RAID group should be the same size. 116 | Physical Storage Management Guide
• The LUNs should be the same size in all RAID groups in the aggregate. 4. Ask the storage array administrator to create the number of LUNs of the size you need for the aggregate. The LUNs should be optimized for performance, according to the instructions in the storage array vendor documentation. 5. Create all the RAID groups in the aggregate at the same time. Note: Do not mix array LUNs from storage arrays with different characteristics in the same Data ONTAP RAID group. Note: If you create a new RAID group for an existing aggregate, be sure that the new RAID group is the same size as the other RAID groups in the aggregate, and that the array LUNs are the same size as the LUNs in the other RAID groups in the aggregate. How Data ONTAP works with hot spare disks A hot spare disk is a disk that is assigned to a storage system but is not in use by a RAID group. It does not yet hold data but is ready for use. If a disk failure occurs within a RAID group, Data ONTAP automatically assigns hot spare disks to RAID groups to replace the failed disks. Minimum number of hot spares you should have Having insufficient spares increases the risk of a disk failure with no available spare, resulting in a degraded RAID group. A spare disk is also required to provide important information (a core file) to technical support in case of a controller disruption. MSATA disks, or disks in a multi-disk carrier, should have four hot spares during steady state operation, and you should never allow the number of MSATA hot spares to dip below two. For RAID groups composed of SSDs, you should have at least one spare disk. For all other Data ONTAP disk types, you should have at least one matching or appropriate hot spare available for each kind of disk installed in your storage system. However, having two available hot spares for all disks provides the best protection against disk failure. Having at least two available hot spares provides the following benefits: • When you have two or more hot spares for a data disk, Data ONTAP can put that disk into the maintenance center if needed. Data ONTAP uses the maintenance center to test suspect disks and take offline any disk that shows problems. • Having two hot spares means that when a disk fails, you still have a spare available if another disk fails before you replace the first failed disk. How Data ONTAP uses RAID to protect your data and data availability | 117
A single spare disk can serve as a hot spare for multiple RAID groups. However, if any disk in those RAID groups fails, then no spare is available for any future disk failures, or for a core file, until the spare is replaced. For this reason, having more than one spare is recommended. Related concepts Spare requirements for multi-disk carrier disks on page 28 What disks can be used as hot spares A disk must conform to certain criteria to be used as a hot spare for a particular data disk. For a disk to be used as a hot spare for another disk, it must conform to the following criteria: • It must be either an exact match for the disk it is replacing or an appropriate alternative. • If SyncMirror is in use, the spare must be in the same pool as the disk it is replacing. • The spare must be owned by the same system as the disk it is replacing. What a matching spare is A matching hot spare exactly matches several characteristics of a designated data disk. Understanding what a matching spare is, and how Data ONTAP selects spares, enables you to optimize your spares allocation for your environment. A matching spare is a disk that exactly matches a data disk for all of the following criteria: • Effective Data ONTAP disk type The effective disk type can be affected by the value of the raid.mix.hdd.performance and raid.mix.hdd.capacity options, which determine the disk types that are considered to be equivalent. • Size • Speed (RPM) • Checksum type (BCS or AZCS) Related concepts How Data ONTAP reports disk types on page 10 What an appropriate hot spare is If a disk fails and no hot spare disk that exactly matches the failed disk is available, Data ONTAP uses the best available spare. Understanding how Data ONTAP chooses an appropriate spare when there is no matching spare enables you to optimize your spare allocation for your environment. Data ONTAP picks a non-matching hot spare based on the following criteria: 118 | Physical Storage Management Guide
• If the available hot spares are not the correct size, Data ONTAP uses one that is the next size up, if there is one. The replacement disk is downsized to match the size of the disk it is replacing; the extra capacity is not available. • If the available hot spares are not the correct speed, Data ONTAP uses one that is a different speed. Using drives with different speeds within the same aggregate is not optimal. Replacing a disk with a slower disk can cause performance degradation, and replacing a disk with a faster disk is not cost-effective. • If the failed disk is part of a mirrored aggregate and there are no hot spares available in the correct pool, Data ONTAP uses a spare from the other pool. Using drives from the wrong pool is not optimal because you no longer have fault isolation for your SyncMirror configuration. If no spare exists with an equivalent disk type or checksum type, the RAID group that contains the failed disk goes into degraded mode; Data ONTAP does not combine effective disk types or checksum types within a RAID group. Related concepts How Data ONTAP reports disk types on page 10 About degraded mode When a disk fails, Data ONTAP can continue to serve data, but it must reconstruct the data from the failed disk using RAID parity. When this happens, the affected RAID group is said to be in degraded mode. The performance of a storage system with one or more RAID groups in degraded mode is decreased. A RAID group goes into degraded mode in the following scenarios: • A single disk fails in a RAID4 group. After the failed disk is reconstructed to a spare, the RAID group returns to normal mode. • One or two disks fail in a RAID-DP group. If two disks have failed in a RAID-DP group, the RAID group goes into double-degraded mode. • A disk is taken offline by Data ONTAP. After the offline disk is brought back online, the RAID group returns to normal mode. Note: If another disk fails in a RAID-DP group in double-degraded mode or a RAID4 group in degraded mode, data loss could occur (unless the data is mirrored). For this reason, always minimize the amount of time a RAID group is in degraded mode by ensuring that appropriate hot spares are available. How Data ONTAP uses RAID to protect your data and data availability | 119
Related concepts How Data ONTAP handles a failed disk that has no available hot spare on page 121 How low spare warnings can help you manage your spare drives By default, Data ONTAP issues warnings to the console and logs if you have fewer than one hot spare drive that matches the attributes of each drive in your storage system. You can change the threshold value for these warning messages to ensure that your system adheres to best practices. To make sure that you always have two hot spares for every drive (a best practice), you can set the min_spare_count RAID option to 2. Setting the min_spare_count RAID option to 0 disables low spare warnings. You might want to do this if you do not have enough drives to provide hot spares (for example, if your storage system does not support external disk shelves). You can disable the warnings only if the following requirements are met: • Your system has 16 or fewer drives. • You have no RAID groups that use RAID4. Note: You cannot create aggregates that use RAID4 protection while the raid.min_spare_count option is set to 0. If either of these requirements is no longer met after this option has been set to 0, the option is automatically set back to 1. How Data ONTAP handles a failed disk with a hot spare Using an available matching hot spare, Data ONTAP can use RAID to reconstruct the missing data from the failed disk onto the hot spare disk with no data service interruption. If a disk fails and a matching or appropriate spare is available, Data ONTAP performs the following tasks: • Replaces the failed disk with a hot spare disk. If RAID-DP is enabled and a double-disk failure occurs in the RAID group, Data ONTAP replaces each failed disk with a separate spare disk. • In the background, reconstructs the missing data onto the hot spare disk or disks. Note: During reconstruction, the system is in degraded mode, and file service might slow down. • Logs the activity in the /etc/messages file. • Sends an AutoSupport message. Attention: Always replace the failed disks with new hot spare disks as soon as possible, so that hot spare disks are always available in the storage system. 120 | Physical Storage Management Guide
Note: If the available spare disks are not the correct size, Data ONTAP chooses a disk of the next larger size and restricts its capacity to match the size of the disk it is replacing. Related concepts How Data ONTAP handles a failed disk that has no available hot spare on page 121 How Data ONTAP handles a failed disk that has no available hot spare When a failed disk has no appropriate hot spare available, Data ONTAP puts the affected RAID group into degraded mode indefinitely and the storage system automatically shuts down within a specified time period. If the maximum number of disks have failed in a RAID group (two for RAID-DP, one for RAID4), the storage system automatically shuts down in the period of time specified by the raid.timeout option. The default timeout value is 24 hours. To ensure that you are aware of the situation, Data ONTAP sends an AutoSupport message whenever a disk fails. In addition, it logs a warning message in the /etc/message file once per hour after a disk fails. Attention: If a disk fails and no hot spare disk is available, contact technical support. Related concepts How Data ONTAP handles a failed disk with a hot spare on page 120 About degraded mode on page 119 Considerations for changing the timeout RAID option The raid.timeout option controls how long a storage system runs after a RAID group goes into degraded mode or the NVRAM battery malfunctions or loses power. You can change the value of this option, but you should understand the implications of doing so. The purpose for the system shutdown is to avoid data loss, which can happen if an additional disk failure occurs in a RAID group that is already running in degraded mode, or if a stand-alone system encounters a catastrophic error and has to shut down without NVRAM. You can extend the number of hours the system operates in these conditions by increasing the value of this option (the default value is 24). You can even disable the shutdown by setting the option to 0, but the longer the system operates with one or both of these conditions, the greater the chance of incurring data loss. How Data ONTAP uses RAID to protect your data and data availability | 121
How RAID-level disk scrubs verify data integrity RAID-level scrubbing means checking the disk blocks of all disks in use in aggregates (or in a particular aggregate, plex, or RAID group) for media errors and parity consistency. If Data ONTAP finds media errors or inconsistencies, it uses RAID to reconstruct the data from other disks and rewrites the data. RAID-level scrubs help improve data availability by uncovering and fixing media and checksum errors while the RAID group is in a normal state (for RAID-DP, RAID-level scrubs can also be performed when the RAID group has a single-disk failure). RAID-level scrubs can be scheduled or run manually. Changing the schedule for automatic RAID-level scrubs You can change the start time and duration of the scheduled scrubs if you want your data to be scrubbed more often than the default RAID-level scrub schedule allows. About this task The default RAID-level scrub schedule is to start a scrub every day starting at 1:00 a.m. On Sundays, the scrub runs for twelve hours; all other days it runs for four hours. Steps 1. Update the RAID-level scrub schedule: storage raid-options modify -node nodename -name raid.scrub.schedule - value duration[h|m]@weekday@start_time,[duration[h| m]@weekday@start_time,...] • duration is specified in either hours or minutes. • weekday is the day of the week: mon, tue, wed, thu, fri, sat, or sun. • start_time is the hour when the scrub should start, using 24-hour format. The current automatic RAID-level scrub schedule is replaced by the schedule you specify. 2. Confirm your new schedule: storage raid-options show raid.scrub.schedule Example The following command prevents any RAID-level scrubbing on node sys1-a from occurring on Mondays, and increases the time duration on Saturdays to eight hours: 122 | Physical Storage Management Guide
storage raid-options modify -node sys1-a -name raid.scrub.schedule - value 4h@tue@1,4h@wed@1,4h@thu@1,4h@fri@1,8h@sat@1,12h@sun@1 How you run a manual RAID-level scrub You can manually run a RAID-level scrub on individual RAID groups, plexes, aggregates, or all aggregates using the storage aggregate scrub command. You can also stop, suspend, and resume manual RAID-level scrubs. If you try to run a RAID-level scrub on a RAID group that is not in a normal state (for example, a group that is reconstructing or degraded), the scrub returns errors and does not check that RAID group. You can run a RAID-level scrub on a RAID-DP group with one failed disk. Controlling the impact of RAID operations on system performance You can reduce the impact of RAID operations on system performance by decreasing the speed of the RAID operations. You can control the speed of the following RAID operations with RAID options: • RAID data reconstruction • Disk scrubbing • Plex resynchronization • Synchronous mirror verification The speed that you select for each of these operations might affect the overall performance of the storage system. However, if the operation is already running at the maximum speed possible and it is fully utilizing one of the three system resources (the CPU, disks, or the disk-to-controller connection bandwidth), changing the speed of the operation has no effect on the performance of the operation or the storage system. If the operation is not yet running, you can set a speed that minimally slows storage system network operations or a speed that severely slows storage system network operations. For each operation, use the following guidelines: • If you want to reduce the performance impact on client access to the storage system, change the specific RAID option from medium to low. Doing so also causes the operation to slow down. • If you want to speed up the operation, change the RAID option from medium to high. Doing so might decrease the performance of the storage system in response to client access. How Data ONTAP uses RAID to protect your data and data availability | 123
Controlling the performance impact of RAID data reconstruction Because RAID data reconstruction consumes CPU resources, increasing the speed of data reconstruction sometimes slows storage system network and disk operations. You can control the speed of data reconstruction with the raid.reconstruc.perf_impact option. About this task When RAID data reconstruction and plex resynchronization are running at the same time, Data ONTAP limits the combined resource utilization to the greater impact set by either operation. For example, if raid.resync.perf_impact is set to medium and raid.reconstruct.perf_impact is set to low, the resource utilization of both operations has a medium impact. The setting for this option also controls the speed of Rapid RAID Recovery. Step 1. Enter the following command: storage raid-options modify -node node_name raid.reconstruct.perf_impact impact impact can be high, medium, or low. high means that the storage system uses most of the system resources available for RAID data reconstruction; this setting can heavily affect storage system performance, but reconstruction finishes sooner, reducing the time that the RAID group is in degraded mode. low means that the storage system uses very little of the system resources; this setting lightly affects storage system performance. However, reconstruction takes longer to complete, increasing the time that the storage system is running in degraded mode. The default impact is medium. Controlling the performance impact of RAID-level scrubbing When Data ONTAP performs a RAID-level scrub, it checks the disk blocks of all disks on the storage system for media errors and parity consistency. You can control the impact this operation has on system performance with the raid.verify.perf_impact option. About this task When RAID-level scrubbing and mirror verification are running at the same time, Data ONTAP limits the combined resource utilization to the greater impact set by either operation. For example, if raid.verify.perf_impact is set to medium and raid.scrub.perf_impact is set to low, the resource utilization by both operations has a medium impact. 124 | Physical Storage Management Guide
If there are times during the day when the load on your storage system is decreased, you can also limit the performance impact of the automatic RAID-level scrub by changing the start time or duration of the automatic scrub. Step 1. Enter the following command: storage raid-options modify -node node_name raid.scrub.perf_impact impact impact can be high, medium, or low. high means that the storage system uses most of the system resources available for scrubbing; this setting can heavily affect storage system performance, but the scrub finishes sooner. low means that the storage system uses very little of the system resources; this setting lightly affects storage system performance, but the scrub takes longer to complete. The default impact is low. Controlling the performance impact of plex resynchronization Plex resynchronization ensures that both plexes of a mirrored aggregate are identical. You can control the performance impact of plex resynchronization by using the raid.resync.perf_impact option. About this task When plex resynchronization and RAID data reconstruction are running at the same time, Data ONTAP limits the combined resource utilization to the greatest impact set by either operation. For example, if raid.resync.perf_impact is set to medium and raid.reconstruct.perf_impact is set to low, the resource utilization by both operations has a medium impact. You should set this option to the same value for both nodes in an HA configuration. Step 1. Enter the following command: storage raid-options modify -node node_name raid.resync.perf_impact impact impact can be high, medium, or low. high means that the storage system uses most of the system resources available for plex resynchronization; this setting can heavily affect storage system performance, but the resynchronization finishes sooner. low means that the storage system uses very little of the system resources; this setting lightly affects storage system performance, but the resynchronization takes longer to complete. How Data ONTAP uses RAID to protect your data and data availability | 125
The default impact is medium. Controlling the performance impact of mirror verification You use mirror verification to ensure that the two plexes of a synchronous mirrored aggregate are identical. You can control the speed of mirror verification, and its effect on system resources, by using the raid.verify.perf_impact option. About this task When mirror verification and RAID-level scrubbing are running at the same time, Data ONTAP limits the combined resource utilization to the greatest impact set by either operation. For example, if raid.verify.perf_impact is set to medium and raid.scrub.perf_impact is set to low, the resource utilization of both operations has a medium impact. For more information about synchronous mirroring, see the Clustered Data ONTAP Data Protection Tape Backup and Recovery Guide. Step 1. Enter the following command: storage raid-options modify -node node_name raid.verify.perf_impact impact impact can be high, medium, or low. high means that the storage system uses most of the system resources available for mirror verification; this setting can heavily affect storage system performance, but the mirror verification finishes sooner. low means that the storage system uses very little of the system resources; this setting lightly affects storage system performance, but the mirror verification takes longer to complete. The default impact is low. 126 | Physical Storage Management Guide
What aggregates are To support the differing security, backup, performance, and data sharing needs of your users, you can group the physical data storage resources on your storage system into one or more aggregates. You can then design and configure these aggregates to provide the appropriate level of performance and redundancy. Each aggregate has its own RAID configuration, plex structure, and set of assigned drives or array LUNs. The aggregate provides storage, based on its configuration, to its associated FlexVol volumes or Infinite Volume. Aggregates have the following characteristics: • They can be composed of drives or array LUNs. • They can be mirrored or unmirrored. • If they are composed of drives, they can be single-tier (composed of only HDDs or only SSDs) or they can be Flash Pool aggregates, which include both HDD RAID groups and an SSD cache. The cluster administrator can assign one or more aggregates to a Storage Virtual Machine (SVM), in which case you can use only those aggregates to contain volumes for that SVM. Related information NetApp Technical Report 3437: Storage Subsystem Resiliency Guide How unmirrored aggregates work Unless you are using SyncMirror, all of your aggregates are unmirrored. Unmirrored aggregates have only one plex (copy of their data), which contains all of the RAID groups belonging to that aggregate. The following diagram shows an unmirrored aggregate composed of disks, with its one plex. The aggregate has four RAID groups: rg0, rg1, rg2, and rg3. Each RAID group has 6 data disks, one parity disk, and one dparity (double parity) disk. All disks used by the aggregate come from the same pool, pool0. 127
Spare disk Data disk Parity disk dParity disk RAID group Aggregate Plex0 (pool0) pool0 rg0 rg1 rg2 rg3 Legend The following diagram shows an unmirrored aggregate with array LUNs, with its one plex. It has two RAID groups, rg0 and rg1. All array LUNs used by the aggregate come from the same pool, pool0. Aggregate Plex0 (pool0) rg0 rg1 Legend array LUN in the aggregate Data ONTAP RAID group 128 | Physical Storage Management Guide
How mirrored aggregates work Mirrored aggregates have two plexes (copies of their data), which use the SyncMirror functionality to duplicate the data to provide redundancy. When a mirrored aggregate is created (or when a second plex is added to an existing unmirrored aggregate), Data ONTAP copies the data in the original plex (plex0) to the new plex (plex1). The plexes are physically separated (each plex has its own RAID groups and its own pool), and the plexes are updated simultaneously. This provides added protection against data loss if more disks fail than the RAID level of the aggregate protects against or there is a loss of connectivity, because the unaffected plex continues to serve data while you fix the cause of the failure. After the plex that had a problem is fixed, the two plexes resynchronize and reestablish the mirror relationship. Note: The time for the two plexes to resynchronize can vary and depends on many variables such as aggregate size, system load, how much data has changed, and so on. The disks and array LUNs on the system are divided into two pools: pool0 and pool1. Plex0 gets its storage from pool0 and plex1 gets its storage from pool1. The following diagram shows an aggregate composed of disks with SyncMirror enabled and implemented. A second plex has been created for the aggregate, plex1. The data in plex1 is a copy of the data in plex0, and the RAID groups are also identical. The 32 spare disks are allocated to pool0 or pool1, 16 disks for each pool. Aggregate Plex0 (pool0) Plex1 (pool1) pool0 pool1 rg0 rg1 rg2 rg3 rg0 rg1 rg2 rg3 Spare disk Data disk Parity disk dParity disk RAID group Legend What aggregates are | 129
The following diagram shows an aggregate composed of array LUNs with SyncMirror enabled and implemented. A second plex has been created for the aggregate, plex1. Plex1 is a copy of plex0, and the RAID groups are also identical. Aggregate Plex0 (pool0) Plex1 (pool1) rg0rg0 rg1rg1 array LUN in the aggregate Data ONTAP RAID group What a Flash Pool aggregate is A Flash Pool aggregate combines both SSDs and HDDs (performance or capacity) to provide a high- performance aggregate more economically than an SSD aggregate. The SSDs provide a high-performance cache for the active data set of the data volumes provisioned on the Flash Pool aggregate, offloading random read operations and repetitive random write operations to improve response times and overall throughput for disk I/O-bound data access operations. (Performance is not significantly increased for predominately sequential workloads.) Related tasks Creating a Flash Pool aggregate using physical SSDs on page 171 How Flash Pool aggregates work In general, Flash Pool aggregates are used and managed in a similar manner as standard aggregates. However, you need to understand how both the SSD and HDD RAID groups interact and affect the rest of the system. The SSD RAID groups, also called the SSD cache, can be composed of physical SSDs or allocation units from SSD storage pools (but not both). The SSD cache does not contribute to the size of the aggregate as calculated against the maximum aggregate size. For example, even if an aggregate is at the maximum aggregate size, you can add an SSD RAID group to it. The SSDs do count toward the overall (node or HA pair) drive limit. 130 | Physical Storage Management Guide
The HDD RAID groups in a Flash Pool aggregate behave the same as HDD RAID groups in a standard aggregate, following the same rules for mixing disk types, sizes, speeds, and checksums. For example, you cannot combine performance and capacity disks in the HDD RAID groups of a Flash Pool aggregate. The checksum type, RAID type, and RAID group size values can be configured for the SSD cache RAID groups and HDD RAID groups independently. If the Flash Pool aggregate uses an SSD storage pool for its SSD cache, the cache RAID type can be changed only when the first SSD RAID groups are added, and the size of the SSD RAID groups are determined by the number of SSDs in the storage pool. There is a platform-dependent maximum size for the SSD cache. Related concepts Rules for mixing drive types in Flash Pool aggregates on page 147 Rules for mixing HDD types in aggregates on page 146 How the available Flash Pool cache capacity is calculated on page 133 Related tasks Changing the RAID type of RAID groups in a Flash Pool aggregate on page 180 Related information NetApp Hardware Universe Requirements for using Flash Pool aggregates The Flash Pool technology has some configuration requirements that you should be aware of before planning to use it in your storage architecture. Flash Pool aggregates cannot be used in the following configurations: You cannot use the following types of storage objects in Flash Pool aggregates: • SSDs that are partitioned for root-data partitioning Partitioned HDDs can be used for the HDD RAID groups of a Flash Pool aggregate. • Array LUNs You can use Flash Pool aggregates and the Flash Cache module (WAFL external cache) in the same system. However, data stored in a Flash Pool aggregate is not cached in the Flash Cache module. Flash Cache is reserved for data stored in aggregates composed of only HDDs. You can use data compression on volumes associated with a Flash Pool aggregate. However, compressed blocks are cached in the Flash Pool cache only for read operations; compressed blocks are not cached for write operations. What aggregates are | 131
Data in read-only volumes, such as SnapMirror or SnapVault destinations, is cached in the Flash Pool cache. Flash Pool aggregates can be created from mirrored aggregates; however, the SSD configuration must be kept the same for both plexes. If you create a Flash Pool aggregate using an aggregate that was created using Data ONTAP 7.1 or earlier, the volumes associated with that Flash Pool aggregate do not support write caching. Related information NetApp Hardware Universe NetApp Technical Report 4070: Flash Pool Design and Implementation Guide Considerations for RAID type and spare management for Flash Pool cache Because of the higher cost of SSDs, it is especially important to find the right balance between cost and protection against drive failure. Knowing how the performance and availability of Flash Pool aggregates are affected by SSDs becoming unavailable can help you determine the correct approach for your needs. The failure or unavailability of an SSD affects its RAID group and aggregate the same way as for HDDs. When an SSD being used in a Flash Pool cache becomes unavailable, the RAID group that contains that SSD goes into degraded mode, and the cache performance for that Flash Pool aggregate is reduced until the RAID group can be reconstructed by copying the contents of the unavailable SSD to a spare SSD. As with HDD RAID groups, if a cache RAID group experiences more concurrent SSD failures than its RAID level can correct for (3 for RAID-DP, 2 for RAID4 ), data integrity and availability can be compromised. Therefore, for maximum data availability, you should use RAID-DP for both the HDD RAID groups and the cache. In addition, you should have a spare SSD available at all times so that reconstruction can begin immediately in the case of an SSD failure. The spare SSD can be shared between multiple Flash Pool aggregates, although all Flash Pool aggregates using that spare will be without a spare after an SSD fails until the spare can be replaced. If using RAID-DP and providing a spare for the Flash Pool cache is prohibitively expensive, and if your RAID group size is not larger than the maximum RAID4 RAID group size, the next best alternative is to use RAID4 for the Flash Pool cache (you should still use RAID-DP for the HDD RAID groups), and provide a spare SSD. This is preferable to using RAID-DP with no spare, because double SSD failures are relatively rare, especially with a smaller RAID group size, and reconstruction takes less time than procuring and installing a spare. Note: If you are using SSD storage pools, you cannot change the RAID type of the cache RAID groups after they have been added to an aggregate. If you believe that you might want to increase the size of your storage pools over time, you should consider the maximum size you will require when determining the correct RAID type to use for the SSD cache. 132 | Physical Storage Management Guide
How Flash Pool aggregates and Flash Cache compare Both the Flash Pool technology and the family of Flash Cache modules (Flash Cache and Flash Cache 2) provide a high-performance cache to increase storage performance. However, there are differences between the two technologies that you should understand before choosing between them. You can employ both technologies on the same system. However, data stored in volumes associated with a Flash Pool aggregate is not cached by Flash Cache. Criteria Flash Pool aggregate Flash Cache Scope A specific aggregate All aggregates assigned to a node Caching types supported Read and write Read Cached data availability during and after takeover events Cached data is available and unaffected by either planned or unplanned takeover events. Cached data is not available during takeover events. After giveback for a planned takeover, previously cached data that is still valid is re- cached automatically. PCIe slot on storage controller required? No Yes Compressed blocks cached? Yes No Supported with array LUNs? No Yes Supported with Storage Encryption? Yes Yes. Data in the cache is not encrypted. For more information about Flash Cache, see the Clustered Data ONTAP System Administration Guide for Cluster Administrators. How the available Flash Pool cache capacity is calculated Flash Pool cache capacity cannot exceed a platform-dependent limit for the node or HA configuration. Knowing the available cache capacity enables you to determine how much cache capacity you can add to your Flash Pool aggregates before reaching the limit. The current Flash Pool cache capacity is the sum of all of the Flash Pool caches (as reported by the storage aggregate show command). For nodes using SyncMirror, including MetroCluster configurations, only the Flash Pool cache of one plex counts toward the cache limit. If Flash Cache modules are installed in a node, the available cache capacity for Flash Pool use is the Flash Pool cache capacity limit minus the sum of the Flash Cache module cache installed on the node. (In the unusual case where the size of the Flash Cache modules is not symmetrical between the What aggregates are | 133
two nodes in an HA configuration, the available Flash Pool cache capacity is decreased by the size of the larger Flash Cache module.) For nodes in an HA configuration, the cache size limit applies to the HA configuration as a whole, and can be split arbitrarily between the two nodes, provided that the total limit for the HA configuration is not exceeded. For nodes in a MetroCluster configuration, the cache size limit applies to the configuration as a whole, but it cannot be split arbitrarily between the nodes as it can for an HA configuration. This means that the limit for each node in the MetroCluster configuration is one quarter of the total cache size limit. If Flash Cache is present on any of the nodes, then the size of the largest Flash Cache module must be subtracted from the overall limit before the per-node limit is calculated. Calculation with Flash Cache modules For an HA configuration composed of two storage controllers with a maximum cache capacity of 12 TB and 2 TB of Flash Cache installed on each node, the maximum Flash Pool aggregate cache capacity for the HA pair would be 12 TB minus 2 TB, or 10 TB. Calculation with asymmetrically sized Flash Cache modules For an HA configuration composed of two storage controllers with a maximum cache capacity of 12 TB and 2 TB of Flash Cache installed on one node and 3 TB of Flash Cache installed on the other node, the maximum Flash Pool cache capacity for the HA pair would be 12 TB minus 3 TB, or 9 TB. Calculation for a MetroCluster configuration For a MetroCluster configuration composed of four storage controllers with a maximum cache capacity of 12 TB, the maximum Flash Pool cache capacity for each node would be 12 TB divided by four, or 3 TB. Calculation for a MetroCluster configuration with Flash Cache modules For a MetroCluster configuration composed of four storage contollers with a maximum cache capacity of 12 TB and 2 TB of Flash Cache installed on one node, the maximum Flash Pool cache capacity for each node would be 12 TB minus 2 TB divided by four, or 2.5 TB. Related concepts How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates on page 137 How Flash Pool aggregates work on page 130 134 | Physical Storage Management Guide
Related references Commands for managing aggregates on page 193 Related information NetApp Hardware Universe Using caching policies on volumes in Flash Pool aggregates You can change the caching policy for a volume that resides on Flash Pool aggregates by using the - caching-policy parameter in the volume create command. When you create a volume on a Flash Pool aggregate, by default, the auto caching policy is assigned to the volume. In most cases, the default caching policy is preferable. The caching policy for a volume should be changed only if a different policy provides better performance. You can set the caching policies for a volume when you create a volume on the Flash Pool aggregate. You can modify the caching policy by using the volume modify command. The caching policies can also be moved between a Flash Pool aggregate and non-Flash Pool aggregate. The following is the list of caching policies, descriptions, and the combinations of read and write caching policies that you can set based on the usage of the volume: Policy Name Description Read caching policy Write caching policy Privilege auto Read caches all metadata blocks and randomly read user data blocks and write caches all randomly overwritten user data blocks. random_read random- write admin none Does not cache any user data or metadata blocks. none none admin random_read Read caches all metadata blocks and randomly read user data blocks. random_read none advanced noread- random_write Write caches all randomly overwritten user data blocks. none random- write advanced meta Read caches only metadata blocks. meta none advanced What aggregates are | 135
Policy Name Description Read caching policy Write caching policy Privilege meta- random_write Read caches all metadata blocks and write caches all randomly overwritten user data blocks. meta random- write advanced random_read_ write Read caches all metadata, randomly read, and randomly written user data blocks. random_read _write none advanced random_read_ write-random- write Read caches all metadata, randomly read, and randomly written user data blocks. It also write caches randomly overwritten user data blocks. random_read _write random- write advanced Related tasks Determining and enabling volume write-caching eligibility for Flash Pool aggregates on page 178 Related information NetApp Technical Report 4070: Flash Pool Design and Implementation Guide Modifying caching policies You should modify the caching policy of a volume only if another policy is expected to provide better performance. The volume must be on a Flash Pool aggregate. Step 1. Use the volume modify command to change the caching policies of a volume. For information about the parameters that you can use with this command, see the volume modify man page. Example The following example modifies the caching policy of a volume named “vol2” to the policy none: volume modify -volume vol2 -caching-policy none 136 | Physical Storage Management Guide
How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates Flash Pool SSD partitioning enables you to group SSDs together into an SSD storage pool that can be allocated to multiple Flash Pool aggregates. This amortizes the cost of the parity SSDs over more aggregates, increases SSD allocation flexibility, and maximizes SSD performance . The storage pool is associated with an HA pair, and can be composed of SSDs owned by either node in the HA pair. When you add an SSD to a storage pool, it becomes a shared SSD, and it is divided into 4 partitions. Storage from an SSD storage pool is divided into allocation units, each of which represents 25% of the total storage capacity of the storage pool. Allocation units contain one partition from each SSD in the storage pool, and are added to a Flash Pool cache as a single RAID group. By default, for storage pools associated with an HA pair, two allocation units are assigned to each of the HA partners, but you can reassign the allocation units to the other HA partner if needed (allocation units must be owned by the node that owns the aggregate). SSD storage pools do not have a RAID type. When an allocation unit is added to a Flash Pool aggregate, the appropriate number of partitions are designated to provide parity to that RAID group. The following diagram shows one example of Flash Pool SSD partitioning. The SSD storage pool pictured is providing cache to two Flash Pool aggregates: What aggregates are | 137
Storage pool SP1 is composed of 5 SSDs; in addition, there is one hot spare SSD available to replace any SSD that experiences a failure. Two of the storage pool's allocation units are allocated to Flash Pool FP1, and two are allocated to Flash Pool FP2. FP1 has a cache RAID type of RAID4, so the allocation units provided to FP1 contain only one partition designated for parity. FP2 has a cache RAID type of RAID-DP, so the allocation units provided to FP2 include a parity partition and a double-parity partition. In this example, two allocation units are allocated to each Flash Pool aggregate; however, if one Flash Pool aggregate needed a larger cache, you could allocate three of the allocation units to that Flash Pool aggregate, and only one to the other. Related concepts What a Flash Pool aggregate is on page 130 Related tasks Creating a Flash Pool aggregate using SSD storage pools on page 173 Considerations for when to use SSD storage pools SSD storage pools provide many benefits, but they also introduce some restrictions that you should be aware of when deciding whether to use SSD storage pools or dedicated SSDs. SSD storage pools make sense only when they are providing cache to two or more Flash Pool aggregates. SSD storage pools provide the following benefits: • Increased storage utilization for SSDs used in Flash Pool aggregates SSD storage pools reduce the overall percentage of SSDs needed for parity by enabling you to share parity SSDs between two or more Flash Pool aggregates. • Ability to share spares between HA partners Because the storage pool is effectively owned by the HA pair, one spare, owned by one of the HA partners, can function as a spare for the entire SSD storage pool if needed. • Better utilization of SSD performance The high performance provided by SSDs can support access by both controllers in an HA pair. These advantages must be weighed against the costs of using SSD storage pools, which include the following items: • Reduced fault isolation The loss of a single SSD affects all RAID groups that include one of its partitions. In this situation, every Flash Pool aggregate that has cache allocated from the SSD storage pool that contains the affected SSD has one or more RAID groups in reconstruction. • Reduced performance isolation 138 | Physical Storage Management Guide
If the Flash Pool cache is not properly sized, there can be contention for the cache between the Flash Pool aggregates that are sharing it. This risk can be mitigated with proper cache sizing and QoS controls. • Decreased management flexibility When you add storage to a storage pool, you increase the size of all Flash Pool caches that include one or more allocation units from that storage pool; you cannot determine how the extra capacity is distributed. How you use SSD storage pools To enable SSDs to be shared by multiple Flash Pool aggregates, you place them in a storage pool. After you add an SSD to a storage pool, you can no longer manage it as a stand-alone entity—you must use the storage pool to assign or allocate the storage provided by the SSD. You create storage pools for a specific HA pair. Then, you add allocation units from that storage pool to one or more Flash Pool aggregates owned by the same HA pair. Just as disks must be owned by the same node that owns an aggregate before they can be allocated to it, storage pools can provide storage only to Flash Pool aggregates owned by one of the nodes that owns the storage pool. If you need to increase the amount of Flash Pool cache on your system, you can add more SSDs to a storage pool, up to the maximum RAID group size for the RAID type of the Flash Pool caches using the storage pool. When you add an SSD to an existing storage pool, you increase the size of the storage pool's allocation units, including any allocation units that are already allocated to a Flash Pool aggregate. You should provide one or more spare SSDs for your storage pools, so that if an SSD in that storage pool becomes unavailable, Data ONTAP can use a spare SSD to reconstruct the partitions of the malfunctioning SSD. You do not need to reserve any allocation units as spare capacity; Data ONTAP can use only a full, unpartitioned SSD as a spare for SSDs in a storage pool. After you add an SSD to a storage pool, you cannot remove it, just as you cannot remove disks from an aggregate. If you want to use the SSDs in a storage pool as discrete drives again, you must destroy all Flash Pool aggregates to which the storage pool's allocation units have been allocated, and then destroy the storage pool. Requirements and best practices for using SSD storage pools There are some technologies that cannot be combined with Flash Pool aggregates that use SSD storage pools. You cannot use the following technologies with Flash Pool aggregates that use SSD storage pools for their cache storage: • MetroCluster • SyncMirror Mirrored aggregates can coexist with Flash Pool aggregates that use storage pools; however, Flash Pool aggregates cannot be mirrored. What aggregates are | 139
• Physical SSDs Flash Pool aggregates can use SSD storage pools or physical SSDs, but not both. SSD storage pools must conform to the following rules: • SSD storage pools can contain only SSDs; HDDs cannot be added to an SSD storage pool. • SSD storage pools can contain between 3 and 28 SSDs. If an SSD storage pool contains more SSDs than the maximum RAID4 RAID group size for SSDs, then it cannot be used for a Flash Pool aggregate whose cache has a RAID type of RAID4. • All SSDs in an SSD storage pool must be owned by the same HA pair. • You cannot use SSDs that have been partitioned for root-data partitioning in a storage pool. If you provide storage from a single storage pool to two caches with different RAID types, and you expand the size of the storage pool beyond the maximum RAID group size for RAID4, the extra partitions in the RAID4 allocation units go unused. For this reason, it is a best practice to keep your cache RAID types homogenous for a storage pool. You cannot change the RAID type of cache RAID groups allocated from a storage pool. You set the RAID type for the cache before adding the first allocation units, and you cannot change it later. When you create a storage pool or add SSDs to an existing storage pool, you must use the same size SSDs. If a failure occurs and no spare of the correct size exists, Data ONTAP can use a larger SSD to replace the failed SSD. However, the larger SSD is right-sized to match the size of the other SSDs in the storage pool, resulting in lost SSD capacity. You can use only one spare SSD for a storage pool. If the storage pool provides allocation units to Flash Pool aggregates owned by both nodes in the HA pair, then the spare SSD can be owned by either node. However, if the storage pool provides allocation units only to Flash Pool aggregates owned by one of the nodes in the HA pair, then the SSD spare must be owned by that same node. Creating an SSD storage pool You create SSD storage pools to provide SSD cache for two to four Flash Pool aggregates. About this task Storage pools do not support a diskcount parameter; you must supply a disk list when creating or adding disks to a storage pool. Steps 1. Determine the names of the spare SSDs available to you: storage aggregate show-spare-disks -disk-type SSD The SSDs used in a storage pool can be owned by either node of an HA pair. 2. Create the storage pool: 140 | Physical Storage Management Guide
storage pool create -storage-pool sp_name -disk-list disk1,disk2,... 3. Optional: Show the newly created storage pool: storage pool show -storage-pool sp_name Result After the SSDs are placed into the storage pool, they no longer appear as spares on the cluster, even though the storage provided by the storage pool has not yet been allocated to any Flash Pool caches. The SSDs can no longer be added to a RAID group as a discrete drive; their storage can be provisioned only by using the allocation units of the storage pool to which they belong. Related concepts How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates on page 137 Considerations for adding SSDs to an existing storage pool versus creating a new one on page 142 Related tasks Creating a Flash Pool aggregate using SSD storage pools on page 173 Adding SSDs to an SSD storage pool on page 141 Related references Commands for managing SSD storage pools on page 143 Adding SSDs to an SSD storage pool When you add SSDs to an SSD storage pool, you increase the storage pool's physical and usable sizes and allocation unit size. The larger allocation unit size also affects allocation units that have already been allocated to Flash Pool aggregates. Before you begin You must have determined that this operation will not cause you to exceed the cache limit for your HA pair. Data ONTAP does not prevent you from exceeding the cache limit when you add SSDs to an SSD storage pool, and doing so can render the newly added storage capacity unavailable for use. About this task When you add SSDs to an existing SSD storage pool, the SSDs must be owned by one node or the other of the same HA pair that already owned the existing SSDs in the storage pool. You can add SSDs that are owned by either node of the HA pair. Steps 1. Optional: View the current allocation unit size and available storage for the storage pool: What aggregates are | 141
storage pool show -instance sp_name 2. Find available SSDs: storage disk show -container-type spare -type SSD 3. Add the SSDs to the storage pool: storage pool add -storage-pool sp_name -disk-list disk1,disk2… The system displays which Flash Pool aggregates will have their size increased by this operation and by how much, and prompts you to confirm the operation. Related concepts How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates on page 137 Considerations for adding SSDs to an existing storage pool versus creating a new one on page 142 How the available Flash Pool cache capacity is calculated on page 133 Related tasks Determining the impact to cache size of adding SSDs to an SSD storage pool on page 143 Creating an SSD storage pool on page 140 Related information NetApp Hardware Universe Considerations for adding SSDs to an existing storage pool versus creating a new one You can increase the size of your SSD cache in two ways—by adding SSDs to an existing SSD storage pool or by creating a new SSD storage pool. The best method for you depends on your configuration and plans for the storage. The choice between creating a new storage pool or adding storage capacity to an existing one is similar to deciding whether to create a new RAID group or add storage to an existing one: • If you are adding a large number of SSDs, creating a new storage pool provides more flexibility because you can allocate the new storage pool differently from the existing one. • If you are adding only a few SSDs, and increasing the RAID group size of your existing Flash Pool caches is not an issue, then adding SSDs to the existing storage pool keeps your spare and parity costs lower, and automatically allocates the new storage. If your storage pool is providing allocation units to Flash Pool aggregates whose caches have different RAID types, and you expand the size of the storage pool beyond the maximum RAID4 RAID group size, the newly added partitions in the RAID4 allocation units are unused. 142 | Physical Storage Management Guide
Determining the impact to cache size of adding SSDs to an SSD storage pool If adding SSDs to a storage pool causes your platform model's cache limit to be exceeded, Data ONTAP does not allocate the newly added capacity to any Flash Pool aggregates. This can result in some or all of the newly added capacity being unavailable for use. About this task When you add SSDs to an SSD storage pool that has allocation units already allocated to Flash Pool aggregates, you increase the cache size of each of those aggregates and the total cache on the system. If none of the storage pool's allocation units have been allocated, adding SSDs to that storage pool does not affect the SSD cache size until one or more allocation units are allocated to a cache. Steps 1. Determine the usable size of the SSDs you are adding to the storage pool: storage disk show disk_name -fields usable-size 2. Determine how many allocation units remain unallocated for the storage pool: storage pool show-available-capacity sp_name All unallocated allocation units in the storage pool are displayed. 3. Calculate the amount of cache that will be added by applying the following formula: ( 4 – number of unallocated allocation units) × 25% × usable size × number of SSDs Commands for managing SSD storage pools Data ONTAP provides the storage pool command for managing SSD storage pools. If you want to... Use this command... Display how much storage a storage pool is providing to which aggregates storage pool show-aggregate Display how much cache would be added to the overall cache capacity for both RAID types (allocation unit data size) storage pool show -instance Display the disks in a storage pool storage pool show-disks Display the unallocated allocation units for a storage pool storage pool show-available- capacity What aggregates are | 143
If you want to... Use this command... Change the ownership of one or more allocation units of a storage pool from one HA partner to the other storage pool reassign Related information Clustered Data ONTAP 8.3 Commands: Manual Page Reference How the SVM affects which aggregates can be associated with a FlexVol volume FlexVol volumes are always associated with one Storage Virtual Machine (SVM), and one aggregate that supplies its storage. The SVM can limit which aggregates can be associated with that volume, depending on how the SVM is configured. When you create a FlexVol volume, you specify which SVM the volume will be created on, and which aggregate that volume will get its storage from. All of the storage for the newly created FlexVol volume comes from that associated aggregate. If the SVM for that volume has aggregates assigned to it, then you can use only one of those assigned aggregates to provide storage to volumes on that SVM. This can help you ensure that your SVMs are not sharing physical storage resources inappropriately. This segregation can be important in a multi- tenancy environment, because for some space management configurations, volumes that share the same aggregate can affect each other's access to free space when space is constrained for the aggregate. Aggregate assignment requirements apply to both cluster administrators and SVM administrators. Volume move and volume copy operations are not constrained by the SVM aggregate assignments, so if you are trying to keep your SVMs on separate aggregates, you must ensure that you do not violate your SVM aggregate assignments when you perform those operations. If the SVM for that volume has no aggregates assigned to it, then the cluster administrator can use any aggregate in the cluster to provide storage to the new volume. However, the SVM administrator cannot create volumes for SVMs with no assigned aggregates. For this reason, if you want your SVM administrator to be able to create volumes for a specific SVM, then you must assign aggregates to that SVM. Changing the aggregates assigned to an SVM does not affect any existing volumes. For this reason, the list of aggregates assigned to an SVM cannot be used to determine the aggregates associated with volumes for that SVM. Related tasks Assigning aggregates to SVMs on page 190 144 | Physical Storage Management Guide
Related information Clustered Data ONTAP 8.3 System Administration Guide for Cluster Administrators Understanding how Data ONTAP works with heterogeneous storage When you have disks with different characteristics (type, speed, size, checksum) or have both disks and array LUNs attached to your storage system, you have heterogeneous storage. Understanding how Data ONTAP works with heterogeneous storage helps you ensure that your aggregates and RAID groups follow best practices and provide maximum storage availability. How you can use disks with mixed speeds in the same aggregate Whenever possible, you should use disks of the same speed in an aggregate. However, if needed, you can configure Data ONTAP to allow mixed speed aggregates based on the disk class. To configure Data ONTAP to allow mixed speed aggregates, you use the following RAID options: • raid.mix.hdd.rpm.performance • raid.mix.hdd.rpm.capacity When these options are set to on, Data ONTAP allows mixing speeds for the designated disk class. Performance disk types are FC and SAS; capacity disk types are BSAS, FSAS, MSATA, and ATA. By default, raid.mix.hdd.rpm.performance is set to off, and raid.mix.hdd.rpm.capacity is set to on. Even if Data ONTAP is not configured to allow mixing speeds, you can still create aggregates out of disks with different speeds by setting the -allow-mixed parameter to true. How to control disk selection from heterogeneous storage When disks with different characteristics coexist on the same node, or when both disks and array LUNs are attached to the same node, the system has heterogeneous storage. When you create an aggregate from heterogeneous storage, you should take steps to ensure that Data ONTAP uses the disks you expect. If your node has heterogeneous storage and you do not explicitly specify what type of disks to use, Data ONTAP uses the disk type (including array LUNs) with the highest number of available disks. When you create or add storage to an aggregate using heterogeneous storage, you should use one of the following methods to ensure that Data ONTAP selects the correct disks or disk types: • Through disk attributes: ◦ You can specify disk size by using the -disksize option. Disks with a usable size between 90% and 105% of the specified size are selected. What aggregates are | 145
◦ You can specify disk speed by using the -diskrpm option. ◦ You can specify disk type by using the -disktype option. • Through disk selection preview. You can use the -simulate option to identify which disks Data ONTAP selects automatically. If you are happy with the disks selected, you can proceed with automatic disk selection. Otherwise, you can use one of the previous methods to ensure that the correct disks are selected. • Through an explicit disk list. You can list the names of specific disks you want to use. However, it is a best practice to use the diskcount option rather than disk lists; this allows Data ONTAP to make the best disk selection for your configuration. Note: For unplanned events such as disk failures, which cause Data ONTAP to add another disk to a RAID group automatically, the best way to ensure that Data ONTAP chooses the best disk for any RAID group on your system is to always have at least one spare (and preferably two) available to match all disk types and sizes in use in your system. Rules for mixing HDD types in aggregates You can mix disks from different loops or stacks within the same aggregate. Depending on the value of the raid.mix.hdd.disktype RAID options, you can mix certain types of HDDs within the same aggregate, but some disk type combinations are more desirable than others. When the appropriate raid.mix.hdd.disktype option is set to off, HDD RAID groups can be composed of only one Data ONTAP disk type. This setting ensures that your aggregates are homogeneous, and requires that you provide sufficient spare disks for every disk type in use in your system. The default value for the raid.mix.hdd.disktype.performance option is off, to prevent mixing SAS and FCAL disks. The default value for the raid.mix.hdd.disktype.capacity option is on. For this setting, the BSAS, FSAS, and ATA disk types are considered to be equivalent for the purposes of creating and adding to aggregates, and for spare management. To maximize aggregate performance and for easier storage administration, you should avoid mixing FC-connected and SAS-connected disks in the same aggregate. This is because of the performance mismatch between FC-connected storage shelves and SAS-connected storage shelves. When you mix these connection types in the same aggregate, the performance of the aggregate is limited by the presence of the FC-connected storage shelves, even though some of the data is being served from the higher-performing SAS-connected storage shelves. MSATA disks cannot be mixed with any other disk type in the same aggregate. If your node uses root-data partitioning, the same mixing rules apply to the partitioned HDDs as to HDDs that are not partitioned. Partitioned HDDs can be mixed with HDDs that are not partitioned in the same aggregate if they are otherwise compatible, but not the same RAID group. Physical 146 | Physical Storage Management Guide
(unpartitioned) HDDs that are added to an existing RAID group that contains partitioned HDDs become partitioned. Disks using Storage Encryption have a Data ONTAP disk type of SAS. However, they cannot be mixed with any other disk type, including SAS disks that are not using Storage Encryption. If any disks on a storage system use Storage Encryption, all of the disks on the storage system (and its high- availability partner node) must use Storage Encryption. Note: If you set a raid.mix.hdd.disktype option to off for a system that already contains aggregates with more than one type of HDD, those aggregates continue to function normally and accept both types of HDDs. However, no other aggregates composed of the specified disk type will accept mixed HDD types as long as that option is set to off. Related concepts How Data ONTAP reports disk types on page 10 Related information NetApp Technical Report 3437: Storage Subsystem Resiliency Guide Rules for mixing drive types in Flash Pool aggregates By definition, Flash Pool aggregates contain more than one drive type. However, the HDD RAID groups follow the same drive-type mixing rules as standard aggregates. For example, you cannot mix performance and capacity disks in the same Flash Pool aggregate. The SSD cache can contain only SSDs. Rules for mixing storage in array LUN aggregates When planning for aggregates, you must consider the rules for mixing storage in aggregates. You cannot mix different storage types or array LUNs from different vendors or vendor families in the same aggregate. Adding the following to the same aggregate is not supported: • Array LUNs and disks • Array LUNs with different checksum types • Array LUNs from different drive types (for example, FC and SATA) or different speeds • Array LUNs from different storage array vendors • Array LUNs from different storage array model families Note: Storage arrays in the same family share the same performance and failover characteristics. For example, members of the same family all perform active-active failover, or they all perform active-passive failover. More than one factor might be used to determine storage array families. What aggregates are | 147
For example, storage arrays with different architectures would be in different families even though other characteristics might be the same. How the checksum type is determined for array LUN aggregates Each Data ONTAP aggregate has a checksum type associated with it. The aggregate checksum type is determined by the checksum type of the array LUNs that are added to it. The checksum type of an aggregate is determined by the checksum type of the first array LUN that is added to the aggregate. The checksum type applies to an entire aggregate (that is, to all volumes in the aggregate). Mixing array LUNs of different checksum types in an aggregate is not supported. • An array LUN of type block must be used with block checksum type aggregates. • An array LUN of type zoned must be used with advanced zoned checksum (AZCS or advanced_zoned) type aggregates. Note: Prior to Data ONTAP 8.1.1, zoned checksum array LUNs were used with ZCS (zoned) type aggregates. Starting with 8.1.1, any new aggregates created with zoned checksum array LUNs are AZCS aggregates. However, you can add zoned checksum array LUNs to existing ZCS aggregates. Before you add array LUNs to an aggregate, you must know the checksum type of the LUNs you want to add, for the following reasons: • You cannot add array LUNs of different checksum types to the same aggregate. • You cannot convert an aggregate from one checksum type to the other. When you create an aggregate you can specify the number of array LUNs to be added, or you can specify the names of the LUNs to be added. If you want to specify a number of array LUNs to be added to the aggregate, the same number or more array LUNs of that checksum type must be available. How to determine space usage in an aggregate You can view space usage by all volumes in one or more aggregates with the aggregate show- space command. This helps you see which volumes are consuming the most space in their containing aggregates so that you can take actions to free more space. The used space in an aggregate is directly affected by the space used in the FlexVol volumes and Infinite Volume constituents it contains. Measures that you take to increase space in a volume also affect space in the aggregate. 148 | Physical Storage Management Guide
When the aggregate is offline, no values are displayed. Only non-zero values are displayed in the command output. However, you can use the -instance parameter to display all possible feature rows regardless of whether they are enabled and using any space. A value of - indicates that there is no data available to display. The following rows are included in the aggregate show-space command output: • Volume Footprints The total of all volume footprints within the aggregate. It includes all of the space that is used or reserved by all data and metadata of all volumes in the containing aggregate. It is also the amount of space that is freed if all volumes in the containing aggregate are destroyed. Infinite Volume constituents appear in the output of space usage commands as if the constituents were FlexVol volumes. • Aggregate Metadata The total file system metadata required by the aggregate, such as allocation bitmaps and inode files. • Snapshot Reserve The amount of space reserved for aggregate Snapshot copies, based on volume size. It is considered used space and is not available to volume or aggregate data or metadata. • Snapshot Reserve Unusable The amount of space originally allocated for aggregate Snapshot reserve that is unavailable for aggregate Snapshot copies because it is being used by volumes associated with the aggregate. Can occur only for aggregates with a non-zero aggregate Snapshot reserve. • Total Used The sum of all space used or reserved in the aggregate by volumes, metadata, or Snapshot copies. • Total Physical Used The amount of space being used for data now (rather than being reserved for future use). Includes space used by aggregate Snapshot copies. There is never a row for Snapshot spill. The following example shows the aggregate show-space command output for an aggregate whose Snapshot reserve is 5%. If the Snapshot reserve was 0, the row would not be displayed. cluster1::> storage aggregate show-space Aggregate : wqa_gx106_aggr1 Feature Used Used% -------------------------------- ---------- ------ Volume Footprints 101.0MB 0% Aggregate Metadata 300KB 0% Snapshot Reserve 5.98GB 5% What aggregates are | 149
Total Used 6.07GB 5% Total Physical Used 34.82KB 0% How you can determine and control a volume's space usage in the aggregate You can determine which FlexVol volumes and Infinite Volume constituents are using the most space in the aggregate and specifically which features within the volume. The volume show- footprint command provides information about a volume's footprint, or its space usage within the containing aggregate. The volume show-footprint command shows details about the space usage of each volume in an aggregate, including offline volumes. This command does not directly correspond to the output of the df command, but instead bridges the gap between the output of the volume show-space and aggregate show-space commands. All percentages are calculated as a percent of aggregate size. Only non-zero values are displayed in the command output. However, you can use the -instance parameter to display all possible feature rows regardless of whether they are enabled and using any space. A value of - indicates that there is no data available to display. Infinite Volume constituents appear in the output of space usage commands as if the constituents were FlexVol volumes. The following example shows the volume show-footprint command output for a volume called testvol: cluster1::> volume show-footprint testvol Vserver : thevs Volume : testvol Feature Used Used% -------------------------------- ---------- ----- Volume Data Footprint 120.6MB 4% Volume Guarantee 1.88GB 71% Flexible Volume Metadata 11.38MB 0% Delayed Frees 1.36MB 0% Total Footprint 2.01GB 76% The following table explains some of the key rows of the output of the volume show-footprint command and what you can do to try to decrease space usage by that feature: 150 | Physical Storage Management Guide
Row/feature name Description/contents of row Some ways to decrease Volume Data Footprint The total amount of space used in the containing aggregate by a volume's data in the active file system and the space used by the volume's Snapshot copies. This row does not include reserved space, so if volumes have reserved files, the volume's total used space in the volume show-space command output can exceed the value in this row. • Deleting data from the volume. • Deleting Snapshot copies from the volume. Volume Guarantee The amount of space reserved by the volume in the aggregate for future writes. The amount of space reserved depends on the guarantee type of the volume. Changing the type of guarantee for the volume to none. This row will go to 0. If you configure your volumes with a volume guarantee of none, you should refer to Technical Report 3965 or 3483 for information about how a volume guarantee of none can affect storage availability. Flexible Volume Metadata The total amount of space used in the aggregate by the volume's metadata files. No direct method to control. Delayed Frees Blocks that Data ONTAP used for performance and cannot be immediately freed. When Data ONTAP frees blocks in a FlexVol volume, this space is not always immediately shown as free in the aggregate because operations to free the space in the aggregate are batched for increased performance. Blocks that are declared free in the FlexVol volume but that are not yet free in the aggregate are called “delayed free blocks” until the associated delayed free blocks are processed. For SnapMirror destinations, this row has a value of 0 and is not displayed. No direct method to control. What aggregates are | 151
Row/feature name Description/contents of row Some ways to decrease File Operation Metadata The total amount of space reserved for file operation metadata. After space is used for file operation metadata, it is not returned as free space to the aggregate, but it is reused by subsequent file operations. No direct method to control. Total Footprint The total amount of space that the volume uses in the aggregate. It is the sum of all of the rows. Any of the methods used to decrease space used by a volume. How Infinite Volumes use aggregates Each Infinite Volume distributes data across multiple aggregates from multiple nodes. By understanding the way that Infinite Volumes use aggregates, you can plan your aggregates in a way that supports the Infinite Volumes that you want. For more information about Infinite Volumes, see the Clustered Data ONTAP Infinite Volumes Management Guide. Aggregate requirements for Infinite Volumes The aggregates that are used by an Infinite Volume should be larger than 100 TB with a minimum of 1.1 TB of available space. If the Infinite Volume uses storage classes, the aggregates must also meet the requirements of the storage class. If an aggregate has less than 1.1 TB of available space, it is not used by the Storage Virtual Machine (SVM) with Infinite Volume. If the Infinite Volume uses storage classes, aggregates must meet the requirements of the storage class to be used. For example, if the storage class is designated to use aggregates of type SAS, aggregates created for that storage class must consist entirely of SAS disks. How FlexVol volumes and Infinite Volumes share aggregates Aggregates can be shared among the volumes in a cluster. Each aggregate can contain multiple FlexVol volumes alongside multiple constituents of Infinite Volumes. When you create an Infinite Volume, constituents of the Infinite Volume are placed on aggregates that are assigned to its containing Storage Virtual Machine (SVM). If the SVM with Infinite Volume includes aggregates that contain FlexVol volumes, one or more of the Infinite Volume's constituents might be placed on aggregates that already include FlexVol volumes, if those aggregates meet the requirements for hosting Infinite Volumes. 152 | Physical Storage Management Guide
Similarly, when you create a FlexVol volume, you can associate that FlexVol volume with an aggregate that is already being used by an Infinite Volume. The following diagram illustrates aggregate sharing in a four-node cluster that includes both FlexVol volumes and an Infinite Volume. The Infinite Volume uses the aggregates aggrA, aggrB, aggrC, aggrD, aggrE, and aggrG even though the aggregates aggrB, aggrC, and aggrG already provide storage to FlexVol volumes. (For clarity, the individual constituents that make up the Infinite Volume are not shown.) Infinite Volume aggrA aggrC aggrB aggrD aggrE aggrF aggrHaggrG Node 1 Node 2 Node 3 Node 4 FV FV FV FV FV FV FV FV FV FV FV How storage classes affect which aggregates can be associated with Infinite Volumes Each storage class definition specifies an aggregate type. When you create an Infinite Volume with a storage class, only the type of aggregate specified for the storage class can supply storage for the volume. You must understand storage class definitions to create aggregates that are appropriate for the storage class. Storage class definitions are available only in OnCommand Workflow Automation. After you understand the aggregate requirements for each storage class, you can use the command-line interface or OnCommand Workflow Automation to create aggregates for storage classes. However, you must use OnCommand Workflow Automation, not the command-line interface, to create an Infinite Volume with one or more storage classes. When you use OnCommand Workflow Automation to create an Infinite Volume with a storage class, OnCommand Workflow Automation automatically filters the aggregates available in the cluster based on the storage class that you want to use. If no aggregates meet the requirements of the storage class, you cannot create an Infinite Volume with that storage class. What aggregates are | 153
How aggregates and nodes are associated with Infinite Volumes The aggregate list of the containing Storage Virtual Machine (SVM) with Infinite Volume determines which aggregates the Infinite Volume uses, as well as who can create an Infinite Volume and which nodes the Infinite Volume uses. The aggregate list can be specified or unspecified, which is represented as a dash ("-"). By default, when a cluster administrator creates any SVM, its aggregate list is unspecified. After the SVM is created, the cluster administrator can specify the aggregate list by using the vserver add- aggregates command. Considerations when choosing to specify the aggregate list or leave it unspecified If you are dedicating an entire cluster to the SVM with Infinite Volume, you can leave the aggregate list of an SVM with Infinite Volume unspecified. In most other situations, you should specify the aggregate list of an SVM with Infinite Volume. Leaving the aggregate list of an SVM with Infinite Volume unspecified has the following outcomes: • Only a cluster administrator can create the Infinite Volume, not an SVM administrator. • When the Infinite Volume is created, it uses all nodes in the cluster. • When the Infinite Volume is created, it can potentially use all of the aggregates in the cluster. How the aggregate list contains candidate aggregates The aggregate list of an SVM with Infinite Volume acts only as a candidate aggregate list for an Infinite Volume. An Infinite Volume uses aggregates according to various factors, including the following requirements: • When an Infinite Volume is created, at least one data constituent is created on at least one aggregate from each node in the aggregate list. • An Infinite Volume uses only the aggregates that it requires to meet the capacity requirements for its specified size. If the assigned aggregates have far greater capacity than the Infinite Volume requires when it is first created, some aggregates in the aggregate list might not contain any Infinite Volume constituents. How the aggregate list determines the nodes An Infinite Volume uses every node that has an aggregate in the aggregate list of an SVM with Infinite Volume. 154 | Physical Storage Management Guide
When changes to the aggregate list take effect Changes to the aggregate list do not have any immediate effect. The aggregate list is used only when the size of an Infinite Volume changes. For example, if you add an aggregate to the aggregate list of an SVM with Infinite Volume, that aggregate is not used until you modify the size of the Infinite Volume. If you add aggregates from a new node to the aggregate list and then resize the Infinite Volume, whether the Infinite Volume uses the aggregates from the new node depends on several variables, including the size of existing constituents and how much the Infinite Volume was increased in size. How the aggregate list can be filtered You can filter the aggregate list for the SVM by using advanced parameters that control which aggregates are used for each type of constituent, such as data constituents. Unlike the aggregate list for the SVM, these aggregate-selection parameters apply only to a single operation. For example, if you use the parameter for data constituent aggregates when you create the Infinite Volume and then resize the Infinite Volume without using the parameter, the Infinite Volume uses the SVM aggregate list. How space is allocated inside a new Infinite Volume Several rules govern how space is allocated to constituents when an Infinite Volume is created. Understanding these rules can help you understand the best practices for configuring aggregates for an Infinite Volume. The following rules govern how space is allocated to constituents when an Infinite Volume is created: 1. The namespace constituent and its mirror copies are created. Before any space is allocated for data, the namespace constituent and namespace mirror constituents are created as big as possible within their maximum sizes. a. The namespace constituent is placed on the aggregate with the most available space on any node that the Infinite Volume uses. b. The first namespace mirror constituent is placed on the aggregate with the next most available space, as long as the aggregate is on a node that meets all of the following conditions: • The node is used by the Infinite Volume. • It does not already contain the namespace constituent. • It is preferably not the partner node in the HA pair of the node that contains the namespace constituent. c. If SnapDiff is enabled, additional namespace mirror constituents are placed on the aggregate with the most available space on each remaining node used by the Infinite Volume. 2. The data capacity is divided equally among the nodes that the Infinite Volume uses. What aggregates are | 155
The data capacity of an Infinite Volume is balanced across nodes. Data capacity is the space remaining from the Infinite Volume's size after deducting the space required by the namespace- related constituents. 3. Within each node, individual data constituents are made as big as possible within a specified maximum. Data constituents are always created as big as they are allowed to be within a specified maximum. Each time that Data ONTAP creates a data constituent, it evaluates all of the aggregates that the Infinite Volume uses on the node and selects the aggregate that has the most available space. Relocating ownership of aggregates used by Infinite Volumes If you want to relocate ownership of aggregates that are used by an Infinite Volume with SnapDiff enabled, you must ensure that the destination node has a namespace mirror constituent, and you must perform the aggregate reallocation in a specific order. Before you begin • You must know whether SnapDiff is enabled on the Infinite Volume. If you do not know, you can use the volume show command with the -fields -enable- snapdiff parameter. • You must know the names of the aggregates on the source and destination nodes. About this task • Follow this procedure only if SnapDiff is enabled on an Infinite Volume that uses the node containing the aggregates that are affected by the ownership change. If SnapDiff is not enabled on the Infinite Volume, do not follow this procedure, because the presence of an Infinite Volume does not affect the aggregate relocation operation. • For more information about SnapDiff and about how Infinite Volumes use aggregates, see the Clustered Data ONTAP Infinite Volumes Management Guide. Steps 1. Determine whether the destination node is already used by the Infinite Volume by using the vserver show command with the -instance parameter. • If the Aggregate List includes an aggregate from the destination node, the Infinite Volume already uses the destination node. • If the Aggregate List does not include an aggregate from the destination node, the Infinite Volume does not currently use the destination node. The destination node must have a namespace mirror constituent before you can relocate aggregate ownership. 156 | Physical Storage Management Guide
2. If the destination node is not currently used by the Infinite Volume, add the destination node to the Infinite Volume. a. Determine the size of the Infinite Volume's namespace constituent by using the volume show command with the -is-constituent true parameter and identifying the size of the constituents with “ns” in their names. b. Identify an aggregate on the destination node that has available space to accommodate a namespace mirror constituent by using the aggregate show command. c. Assign the aggregate from the new node to the Infinite Volume by using the vserver modify command with the -aggr-list parameter. When you specify the aggregate list, you should include all the existing aggregates from existing nodes as well as the new aggregate from the new node. d. Increase the size of the Infinite Volume by using the volume modify command with the - size parameter. Increase the size of the Infinite Volume by an amount that is equal or larger than the size of the namespace constituent. A namespace mirror constituent is created on the destination node. 3. Identify the type of Infinite Volume constituents contained by each aggregate on the source node by using the volume show command with the -vserver and -is-constituent true parameters. Constituents with “data” in their names are data constituents. The constituent with a name ending in “_ns” is the namespace constituent. Constituents with “ns_mirror” in their names are namespace mirror constituents. Example In the following output, aggr3 contains a data constituent, aggr1 contains a namespace mirror constituent, and aggr2 contains a namespace mirror constituent: cluster1::> volume show -vserver vs0 -is-constituent true Vserver Volume Aggregate State Type Size Available Used% --------- ------------ ------------ ---------- ---- ---------- ---------- ----- vs0 repo_vol_1024_data0001 aggr3 online RW 100TB 95TB 5% vs0 repo_vol_1024_data0002 vs_aggr online RW 100TB 95TB 5% vs0 repo_vol_1024_data0003 aggr4 online RW 100TB 95TB 5% ... ... ... vs0 repo_vol_ns aggr1 online RW 10TB 9.5TB 5% vs0 repo_vol_ns_mirror0001 aggr2 online DP 10TB 9.5TB 5% 100 entries were displayed. 4. Perform the aggregate relocation in the following way: What aggregates are | 157
a. Divide the aggregates into the following two categories: • The single aggregate that contains either a namespace constituent or a namespace mirror constituent. It might also contain data constituents. • All the other aggregates on the node that contain data constituents. b. Relocate ownership of all aggregates that do not contain a namespace constituent or namespace mirror constituent. Note: If you want to relocate ownership of the aggregate that contains the namespace constituent without also relocating ownership of all of the aggregates that contain data constituents, you must contact technical support to create a namespace mirror constituent on the source node. c. If the remaining aggregate contains only the namespace constituent or if the remaining aggregate contains more than one type of constituent, relocate ownership of the remaining aggregate. If the remaining aggregate contains only a namespace mirror constituent and no data constituents, you do not need to change ownership of the aggregate. After you finish You can consider contacting technical support to delete any excess namespace mirror constituents that are using space unnecessarily. 158 | Physical Storage Management Guide
Managing aggregates You create and manage your aggregates so that they can provide storage to their associated volumes. Creating an aggregate using unpartitioned drives You create an aggregate to provide storage to one or more FlexVol volumes and Infinite Volumes. Aggregates are a physical storage object; they are associated with a specific node in the cluster. Before you begin • You should know what drives or array LUNs will be used in the new aggregate. • You should have determined the correct RAID group size given the number of drives or array LUNs you are using to create the aggregate and your plans for expanding the aggregate in the future. • If you have multiple drive types in your node (heterogeneous storage), you should understand how you can ensure that the correct drive type is selected. About this task This procedure should not be used to create an aggregate composed of root or data partitions. Drives and array LUNs are owned by a specific node; when you create an aggregate, all drives in that aggregate must be owned by the same node, which becomes the home node for that aggregate. Aggregate names must conform to the following requirements: • They must begin with either a letter or an underscore (_). • They can contain only letters, digits, and underscores. • They can contain 250 or fewer characters. Steps 1. Display a list of available spares: storage aggregate show-spare-disks -original-owner node_name 2. Create the aggregate by using the storage aggregate create command. If you are logged in to the cluster on the cluster management interface, you can create an aggregate on any node in the cluster. To ensure that the aggregate is created on a specific node, use the -node parameter or specify drives that are owned by that node. 159
The following list describes some of the options you can specify: • Aggregate's home node (that is, the node on which the aggregate is located unless the aggregate fails over to the node's storage failover partner) • The number of disks or array LUNs to be added, or a list of specific drives or array LUNs that are to be included in the aggregate For optimal disk selection, use the diskcount parameter rather than a disk list. This allows Data ONTAP to make an optimal disk selection for your configuration. • Checksum style to use for the aggregate • Type of drives to use • Size of drives to use • Drive speed to use • RAID type for RAID groups on the aggregate • Maximum number of drives or array LUNs that can be included in a RAID group • Whether drives with different RPM are allowed 3. Monitor zeroing progress and verify the RAID groups and drives of your new aggregate: storage aggregate show-status aggr_name Example You want to create a storage configuration for a single-node cluster and have 4 shelves of SAS disks totaling 96 drives, with 3 drives used for the root aggregate. You decide to allocate the remaining drives in the following manner: • 2 spares • 91 drives used for data in one aggregate: ◦ 4 RAID groups of 18 drives each ◦ 1 RAID group of 19 drives To accomplish this goal, you use the following steps: 1. Create the first RAID group of 19 disks: storage aggregate create -aggregate n01sas01 -node cl-01 -diskcount 19 -maxraidsize 19 -disktype SAS 2. Change the aggregate's maximum RAID group size to 18: storage aggregate modify -aggregate n01sas01 -maxraidsize 18 160 | Physical Storage Management Guide
3. Add the remaining 54 disks as 4 RAID groups of 18 disks each: storage aggregate add-disks -aggregate n01sas01 -node cl-01 - diskcount 54 -disktype SAS Later, if you added a 24-drive shelf to the node and wanted to add the storage to this aggregate, you could again modify the RAID group size to 24 so that the entire shelf could be accommodated in a single RAID group, provided that a RAID group size of 24 is acceptable for your installation. Related concepts What aggregates are on page 127 Considerations for sizing RAID groups on page 114 Understanding how Data ONTAP works with heterogeneous storage on page 145 Related tasks Creating an aggregate using root-data partitioning on page 161 Creating a Flash Pool aggregate using physical SSDs on page 171 Creating an aggregate using root-data partitioning Typically, root aggregates are created in the factory, and data aggregates are created during initial system setup. However, if you need to create a new data aggregate using partitioned drives, there are some differences in the procedure from creating an aggregate using physical, unpartitioned drives. Before you begin You should know what drives or partitions will be used in the new aggregate. You should have determined the correct RAID group size given the number of drives or partitions you are using to create the aggregate and your plans for expanding the aggregate in the future. About this task Drives and partitions are owned by a specific node; when you create an aggregate, all of the partitions you use for the new aggregate must be owned by the same node, which becomes the home node for the new aggregate. When you provision partitions, you must ensure that you do not leave the node without a disk with both partitions as spare. If you do, and the node experiences a controller disruption, valuable information about the problem (the core file) might not be available to provide to technical support. Aggregate names must conform to the following requirements: • They must begin with either a letter or an underscore (_). Managing aggregates | 161
• They can contain only letters, digits, and underscores. • They can contain 250 or fewer characters. Steps 1. View the list of spare data partitions: storage aggregate show-spare-disks -original-owner node_name The list of disks with at least one spare partition is displayed. Data partitions are shown under Local Data Usable. You must use data partitions when you create a data aggregate. 2. Determine how many partitions you want to use in the aggregate. Remember to leave an appropriate number of spare data partitions. Data ONTAP will not use a root partition as a spare for a data aggregate. 3. Simulate the creation of the aggregate: storage aggregate create -aggregate aggr_name -node node_name -diskcount number_of_partitions -simulate true This enables you to see the result of the aggregate creation without actually provisioning any storage. If any warnings are displayed from the simulated command, you can adjust the command and repeat the simulation. 4. Create the aggregate: storage aggregate create -aggregate aggr_name -node node_name -diskcount number_of_partitions 5. Check zeroing status and verify the composition of the aggregate: storage aggregate show-status aggr_name Related concepts Understanding root-data partitioning on page 30 Considerations for sizing RAID groups on page 114 Related tasks Creating an aggregate using unpartitioned drives on page 159 Correcting misaligned spare partitions on page 169 162 | Physical Storage Management Guide
Increasing the size of an aggregate that uses physical drives You can add disks or array LUNs to an aggregate so that it can provide more storage to its associated volumes. Before you begin • You must understand the requirement to add disks or array LUNs owned by the same system and pool • For aggregates composed of disks, you must understand the following: ◦ Benefits of keeping your RAID groups homogeneous for disk size and speed ◦ Which types of disks can be used together ◦ Checksum rules when disks of more than one checksum type are in use ◦ How to ensure that the correct disks are added to the aggregate (the disk addition operation cannot be undone) ◦ How to add disks to aggregates from heterogeneous storage ◦ Minimum number of disks to add for best performance ◦ Number of hot spares you need to provide for protection against disk failures ◦ Requirements for adding disks from multi-disk carrier disk shelves ◦ Requirement to add storage to both plexes of a mirrored aggregate at the same time to ensure that the plexes are the same size and contain the same disk types ◦ If you are adding cache to a Flash Pool aggregate, the cache limit for your system model and how much cache you are adding towards the limit About this task This procedure should not be used for aggregates composed of root or data partitions. Following these best practices when you add storage to an aggregate optimizes aggregate performance: • Add a complete RAID group at one time. The new RAID group does not have to be exactly the same size as the existing RAID groups, but it should not be less than one half the size of the existing RAID groups. Managing aggregates | 163
• If any small RAID groups exist already, you can bring them up to the size of the other RAID groups, as long as you add at least as many data drives as are already in the RAID group. • Avoid adding a small number of drives to an existing RAID group. Doing so results in the added disks being the target for a disproportionate percentage of new data, causing the new disks to become a performance bottleneck. Steps 1. Verify that appropriate spare disks or array LUNs are available for you to add: storage aggregate show-spare-disks -original-owner node_name For disks, make sure that enough of the spares listed are of the correct type, size, speed, and checksum type for the target RAID group in the aggregate to which you are adding the disks. 2. Add the disks or array LUNs: storage aggregate add-disks -aggregate aggr_name [-raidgroup raid_group_name] disks If you are adding disks with a different checksum than the aggregate, as when creating a Flash Pool aggregate, or if you are adding disks to a mixed checksum aggregate, you must use the - checksumstyle parameter. If you are adding disks to a Flash Pool aggregate, you must use the -disktype parameter to specify the disk type. If you specify the -raidgroup parameter, the storage is added to the RAID group you specify. raid_group_name is the name that Data ONTAP gave to the group—for example, rg0. If you are adding SSDs to the SSD cache of a Flash Pool aggregate, you do not need to specify the RAID group name; the SSD RAID group is selected by default based on the type of the disks you are adding. disks specifies the disks to be added in one of the following ways: • -diskcount, usually further qualified by disk type or checksum type • -disklist disk1 [disk2...] If possible, you should use the diskcount option. Doing so allows Data ONTAP to optimize the disk selection for your configuration. If you are adding disks to a mirrored aggregate and you are specifying disk names, you must also use the -mirror-disklist parameter. Related concepts How to control disk selection from heterogeneous storage on page 145 Considerations for sizing RAID groups on page 114 What a Flash Pool aggregate is on page 130 164 | Physical Storage Management Guide
Rules for mixing drive types in Flash Pool aggregates on page 147 Rules for mixing HDD types in aggregates on page 146 How the available Flash Pool cache capacity is calculated on page 133 What happens when you add storage to an aggregate on page 170 Related tasks Increasing the size of an aggregate that uses root-data partitioning on page 165 Related information NetApp Technical Report 4070: Flash Pool Design and Implementation Guide Increasing the size of an aggregate that uses root-data partitioning When you add storage to an existing aggregate that is using partitioned drives, you should be aware of whether you are adding a partitioned drive or an unpartitioned drive, and the tradeoffs of mixing those types of drives in the same RAID group versus creating a new RAID group. Before you begin • You should understand whether you are adding partitions or unpartitioned drives to the aggregate. • You should know what the RAID group size is for the aggregate you are adding the storage to. About this task When you add storage to an existing aggregate, you can let Data ONTAP choose the RAID group to add the storage to, or you can designate the target RAID group for the added storage, including creating a new RAID group. If an unpartitioned drive is added to a RAID group composed of partitioned drives, the new drive is partitioned, leaving an unused spare partition. If you do not want the new drive to be partitioned, you can add it to a RAID group that contains only unpartitioned (physical) drives. However, partitioning a drive might be preferable to creating a new, small RAID group. Following these best practices when you add storage to an aggregate optimizes aggregate performance: • Add a complete RAID group at one time. The new RAID group does not have to be exactly the same size as the existing RAID groups, but it should not be less than one half the size of the existing RAID groups. • If any small RAID groups exist already, you can bring them up to the size of the other RAID groups, as long as you add at least as many data drives as are already in the RAID group. Managing aggregates | 165
• Avoid adding a small number of drives to an existing RAID group. Doing so results in the added drives being the target for a disproportionate percentage of new data, causing the new drives to become a performance bottleneck. When you provision partitions, you must ensure that you do not leave the node without a drive with both partitions as spare. If you do, and the node experiences a controller disruption, valuable information about the problem (the core file) might not be available to provide to technical support. Steps 1. Show the available spare storage on the system that owns the aggregate: storage aggregate show-spare-disks -original-owner node_name You can use the -is-disk-shared parameter to show only partitioned drives or only unpartitioned drives. 2. Show the current RAID groups for the aggregate: storage aggregate show-status aggr_name 3. Simulate adding the storage to the aggregate: storage aggregate add-disks -aggregate aggr_name -diskcount number_of_disks_or_partitions -simulate true This enables you to see the result of the storage addition without actually provisioning any storage. If any warnings are displayed from the simulated command, you can adjust the command and repeat the simulation. 4. Add the storage to the aggregate: storage aggregate add-disks -aggregate aggr_name -diskcount number_of_disks_or_partitions You can use the -raidgroup parameter if you want to add the storage to a different RAID group than the default. If you are adding partitions to the aggregate, you must use a disk that shows available capacity for the required partition type. For example, if you are adding partitions to a data aggregate (and using a disk list), the disk names you use must show available capacity in the Local Data Usable column. 5. Verify that the storage was added successfully: storage aggregate show-status -aggregate aggr_name 6. Ensure that the node still has at least one drive with both the root partition and the data partition as spare: storage aggregate show-spare-disks -original-owner node_name 166 | Physical Storage Management Guide
If the node does not have a drive with both partitions as spare and it experiences a controller disruption, then valuable information about the problem (the core file) might not be available to provide to technical support. Example: Adding partitioned drives to an aggregate The following example shows that the cl1-s2 node has multiple spare partitions available: cl1-s2::> storage aggregate show-spare-disks -original-owner cl1-s2 -is-disk-shared true Original Owner: cl1-s2 Pool0 Shared HDD Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size Status --------------------------- ----- ------ -------------- -------- -------- -------- -------- 1.0.1 BSAS 7200 block 753.8GB 73.89GB 828.0GB zeroed 1.0.2 BSAS 7200 block 753.8GB 0B 828.0GB zeroed 1.0.3 BSAS 7200 block 753.8GB 0B 828.0GB zeroed 1.0.4 BSAS 7200 block 753.8GB 0B 828.0GB zeroed 1.0.8 BSAS 7200 block 753.8GB 0B 828.0GB zeroed 1.0.9 BSAS 7200 block 753.8GB 0B 828.0GB zeroed 1.0.10 BSAS 7200 block 0B 73.89GB 828.0GB zeroed 2 entries were displayed. The following example shows that the data_1 aggregate is composed of a single RAID group of five partitions: cl1-s2::> storage aggregate show-status -aggregate data_1 Owner Node: cl1-s2 Aggregate: data_1 (online, raid_dp) (block checksums) Plex: /data_1/plex0 (online, normal, active, pool0) RAID Group /data_1/plex0/rg0 (normal, block checksums) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ---------- shared 1.0.10 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.5 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.6 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.11 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.0 0 BSAS 7200 753.8GB 828.0GB (normal) 5 entries were displayed. The following example shows which partitions would be added to the aggregate: cl1-s2::> storage aggregate add-disks data_1 -diskcount 5 -simulate true Addition of disks would succeed for aggregate "data_1" on node "cl1-s2". The following disks would be used to add to the aggregate: 1.0.2, 1.0.3, 1.0.4, 1.0.8, 1.0.9. The following example adds five spare data partitions to the aggregate: cl1-s2::> storage aggregate add-disks data_1 -diskcount 5 The following example shows that the data partitions were successfully added to the aggregate: cl1-s2::> storage aggregate show-status -aggregate data_1 Owner Node: cl1-s2 Aggregate: data_1 (online, raid_dp) (block checksums) Plex: /data_1/plex0 (online, normal, active, pool0) RAID Group /data_1/plex0/rg0 (normal, block checksums) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ---------- Managing aggregates | 167
shared 1.0.10 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.5 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.6 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.11 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.0 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.2 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.3 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.4 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.8 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.9 0 BSAS 7200 753.8GB 828.0GB (normal) 10 entries were displayed. The following example verifies that an entire disk, disk 1.0.1, remains available as a spare: cl1-s2::> storage aggregate show-spare-disks -original-owner cl1-s2 -is-disk-shared true Original Owner: cl1-s2 Pool0 Shared HDD Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size Status --------------------------- ----- ------ -------------- -------- -------- -------- -------- 1.0.1 BSAS 7200 block 753.8GB 73.89GB 828.0GB zeroed 1.0.10 BSAS 7200 block 0B 73.89GB 828.0GB zeroed 2 entries were displayed. Related concepts How to control disk selection from heterogeneous storage on page 145 What a Flash Pool aggregate is on page 130 Considerations for sizing RAID groups on page 114 Rules for mixing drive types in Flash Pool aggregates on page 147 Rules for mixing HDD types in aggregates on page 146 How the available Flash Pool cache capacity is calculated on page 133 What happens when you add storage to an aggregate on page 170 Related tasks Correcting misaligned spare partitions on page 169 Increasing the size of an aggregate that uses physical drives on page 163 Related information NetApp Technical Report 4070: Flash Pool Design and Implementation Guide 168 | Physical Storage Management Guide
Correcting misaligned spare partitions When you add partitioned disks to an aggregate, you must leave a disk with both the root and data partition available as spare for every node. If you do not and your node experiences a disruption, Data ONTAP might not be able to create a core file. Before you begin You must have both a spare data partition and a spare root partition on the same type of disk owned by the same node. Steps 1. Display the spare partitions for the node: storage aggregate show-spare-disks -original-owner node_name Note which disk has a spare data partition (spare_data) and which disk has a spare root partition (spare_root). The spare partition will show a non-zero value under the Local Data Usable or Local Root Usable column. 2. Replace the disk with a spare data partition with the disk with the spare root partition: storage disk replace -disk spare_data -replacement spare_root -action start You can copy the data in either direction; however, copying the root partition takes less time to complete. 3. Monitor the progress of the disk replacement: storage aggregate show-status -aggregate aggr_name 4. After the replacement operation is complete, display the spares again to confirm that you have a full spare disk: storage aggregate show-spare-disks -original-owner node_name You should see a spare disk with usable space under both Local Data Usable and Local Root Usable. Example You display your spare partitions for node c1-01 and see that your spare partitions are not aligned: c1::> storage aggregate show-spare-disks -original-owner c1-01 Original Owner: c1-01 Pool0 Shared HDD Spares Local Local Data Root Physical Managing aggregates | 169
Disk Type RPM Checksum Usable Usable Size --------------------------- ----- ------ -------------- -------- -------- -------- 1.0.1 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.10 BSAS 7200 block 0B 73.89GB 828.0GB You start the disk replacement job: c1::> storage disk replace -disk 1.0.1 -replacement 1.0.10 -action start While you are waiting for the replacement operation to finish, you display the progress of the operation: c1::> storage aggregate show-status -aggregate aggr0_1 Owner Node: c1-01 Aggregate: aggr0_1 (online, raid_dp) (block checksums) Plex: /aggr0_1/plex0 (online, normal, active, pool0) RAID Group /aggr0_1/plex0/rg0 (normal, block checksums) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ---------- shared 1.0.1 0 BSAS 7200 73.89GB 828.0GB (replacing, copy in progress) shared 1.0.10 0 BSAS 7200 73.89GB 828.0GB (copy 63% completed) shared 1.0.0 0 BSAS 7200 73.89GB 828.0GB (normal) shared 1.0.11 0 BSAS 7200 73.89GB 828.0GB (normal) shared 1.0.6 0 BSAS 7200 73.89GB 828.0GB (normal) shared 1.0.5 0 BSAS 7200 73.89GB 828.0GB (normal) After the replacement operation is complete, you confirm that you have a full spare disk: ie2220::> storage aggregate show-spare-disks -original-owner c1-01 Original Owner: c1-01 Pool0 Shared HDD Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size --------------------------- ----- ------ -------------- -------- -------- -------- 1.0.1 BSAS 7200 block 753.8GB 73.89GB 828.0GB What happens when you add storage to an aggregate By default, Data ONTAP adds new drives or array LUNs to the most recently created RAID group until it reaches its maximum size. Then Data ONTAP creates a new RAID group. Alternatively, you can specify a RAID group that you want to add storage to. When you create an aggregate or add storage to an aggregate, Data ONTAP creates new RAID groups as each RAID group is filled with its maximum number of drives or array LUNs. The last RAID group formed might contain fewer drives or array LUNs than the maximum RAID group size for the aggregate. In that case, any storage added to the aggregate is added to the last RAID group until the specified RAID group size is reached. If you increase the RAID group size for an aggregate, new drives or array LUNs are added only to the most recently created RAID group; the previously created RAID groups remain at their current size unless you explicitly add storage to them. 170 | Physical Storage Management Guide
If you add a drive to a RAID group that is larger than the drives already there, the new drive is capacity-limited to be the same size as the other drives. Note: You are advised to keep your RAID groups homogeneous when possible. If needed, you can replace a mismatched drive with a more suitable drive later. Related tasks Increasing the size of an aggregate that uses physical drives on page 163 Increasing the size of an aggregate that uses root-data partitioning on page 165 Creating a Flash Pool aggregate using physical SSDs You create a Flash Pool aggregate by enabling the feature on an existing aggregate composed of HDD RAID groups, and then adding one or more SSD RAID groups to that aggregate. This results in two sets of RAID groups for that aggregate: SSD RAID groups (the SSD cache) and HDD RAID groups. Before you begin • You must have identified a valid aggregate composed of HDDs to convert to a Flash Pool aggregate. • You must have determined write-caching eligibility of the volumes associated with the aggregate, and completed any required steps to resolve eligibility issues. • You must have determined the SSDs you will be adding, and these SSDs must be owned by the node on which you are creating the Flash Pool aggregate. • You must have determined the checksum types of both the SSDs you are adding and the HDDs already in the aggregate. • You must have determined the number of SSDs you are adding and the optimal RAID group size for the SSD RAID groups. Using fewer RAID groups in the SSD cache reduces the number of parity disks required, but larger RAID groups require RAID-DP. • You must have determined the RAID level you want to use for the SSD cache. • You must have determined the maximum cache size for your system and determined that adding SSD cache to your aggregate will not cause you to exceed it. • You must have familiarized yourself with the configuration requirements for Flash Pool aggregates. Managing aggregates | 171
About this task After you add an SSD cache to an aggregate to create a Flash Pool aggregate, you cannot remove the SSD cache to convert the aggregate back to its original configuration. By default, the RAID level of the SSD cache is the same as the RAID level of the HDD RAID groups. You can override this default selection by specifying the raidtype option when you add the first SSD RAID groups. Steps 1. Mark the aggregate as eligible to become a Flash Pool aggregate: storage aggregate modify -aggregate aggr_name -hybrid-enabled true If this step does not succeed, determine write-caching eligibility for the target aggregate. 2. Add the SSDs to the aggregate by using the storage aggregate add command. You can specify the SSDs by ID or by using the diskcount and disktype parameters. If the HDDs and the SSDs do not have the same checksum type, or if the aggregate is a mixed- checksum aggregate, then you must use the checksumstyle parameter to specify the checksum type of the disks you are adding to the aggregate. You can specify a different RAID type for the SSD cache by using the raidtype parameter. If you want the cache RAID group size to be different from the default for the RAID type you are using, you should change it now, by using the -cache-raid-group-size parameter. Related concepts What a Flash Pool aggregate is on page 130 Requirements for using Flash Pool aggregates on page 131 Considerations for RAID type and spare management for Flash Pool cache on page 132 How the available Flash Pool cache capacity is calculated on page 133 Related tasks Creating a Flash Pool aggregate using SSD storage pools on page 173 Determining Flash Pool candidacy and optimal cache size on page 175 Determining and enabling volume write-caching eligibility for Flash Pool aggregates on page 178 Creating an aggregate using unpartitioned drives on page 159 Related information NetApp Technical Report 4070: Flash Pool Design and Implementation Guide 172 | Physical Storage Management Guide
Creating a Flash Pool aggregate using SSD storage pools You create a Flash Pool aggregate with SSD storage pools by enabling the feature on an existing aggregate composed of HDD RAID groups, and then adding one or more SSD storage pool allocation units to that aggregate. Before you begin • You must have identified a valid aggregate composed of HDDs to convert to a Flash Pool aggregate. • You must have determined write-caching eligibility of the volumes associated with the aggregate, and completed any required steps to resolve eligibility issues. • You must have created an SSD storage pool to provide the SSD cache to this Flash Pool aggregate. Any allocation unit from the storage pool that you want to use must be owned by the same node that owns the Flash Pool aggregate. • You must have determined how much cache you want to add to the aggregate. You add cache to the aggregate by allocation units. You can increase the size of the allocation units later by adding SSDs to the storage pool if there is room. • You must have determined the RAID type you want to use for the SSD cache. After you add a cache to the aggregate from SSD storage pools, you cannot change the RAID type of the cache RAID groups. • You must have determined the maximum cache size for your system and determined that adding SSD cache to your aggregate will not cause you to exceed it. You can see the amount of cache that will be added to the total cache size by using the storage pool show command. • You must have familiarized yourself with the configuration requirements for Flash Pool aggregates. About this task If you want the RAID type of the cache to different from that of the HDD RAID groups, you must specify the cache RAID type when you add the SSD capacity. After you add the SSD capacity to the aggregate, you can no longer change the RAID type of the cache. After you add an SSD cache to an aggregate to create a Flash Pool aggregate, you cannot remove the SSD cache to convert the aggregate back to its original configuration. Managing aggregates | 173
Steps 1. Mark the aggregate as eligible to become a Flash Pool aggregate: storage aggregate modify -aggregate aggr_name -hybrid-enabled true If this step does not succeed, determine write-caching eligibility for the target aggregate. 2. Show the available SSD storage pool allocation units: storage pool show-available-capacity 3. Add the SSD capacity to the aggregate: storage aggregate add aggr_name -storage-pool sp_name -allocation-units number_of_units If you want the RAID type of the cache to be different from that of the HDD RAID groups, you must change it when you enter this command by using the raidtype parameter. You do not need to specify a new RAID group; Data ONTAP automatically puts the SSD cache into separate RAID groups from the HDD RAID groups. You cannot set the RAID group size of the cache; it is determined by the number of SSDs in the storage pool. The cache is added to the aggregate and the aggregate is now a Flash Pool aggregate. Each allocation unit added to the aggregate becomes its own RAID group. 4. Optional: Confirm the presence and size of the SSD cache: storage aggregate show aggr_name The size of the cache is listed under Total Hybrid Cache Size. Related concepts How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates on page 137 What a Flash Pool aggregate is on page 130 Requirements for using Flash Pool aggregates on page 131 Considerations for RAID type and spare management for Flash Pool cache on page 132 How the available Flash Pool cache capacity is calculated on page 133 Related tasks Creating a Flash Pool aggregate using physical SSDs on page 171 Creating an aggregate using unpartitioned drives on page 159 Determining Flash Pool candidacy and optimal cache size on page 175 Determining and enabling volume write-caching eligibility for Flash Pool aggregates on page 178 174 | Physical Storage Management Guide
Related information NetApp Technical Report 4070: Flash Pool Design and Implementation Guide Determining Flash Pool candidacy and optimal cache size Before converting an existing aggregate to a Flash Pool aggregate, you can determine whether the aggregate is I/O bound, and what would be the best Flash Pool cache size for your workload and budget. You can also check whether the cache of an existing Flash Pool aggregate is sized correctly. Before you begin You should know approximately when the aggregate you are analyzing experiences its peak load. Steps 1. Enter advanced mode: set advanced 2. If you need to determine whether an existing aggregate would be a good candidate for conversion to a Flash Pool aggregate, determine how busy the disks in the aggregate are during a period of peak load, and how that is affecting latency: statistics show-periodic -object disk:raid_group -instance raid_group_name -counter disk_busy|user_read_latency -interval 1 - iterations 60 You can decide whether reducing latency by adding Flash Pool cache makes sense for this aggregate. Example The following command shows the statistics for the first RAID group of the aggregate “aggr1”: statistics show-periodic -object disk:raid_group -instance /aggr1/ plex0/rg0 -counter disk_busy|user_read_latency -interval 1 -iterations 60 3. Start AWA: system node run -node node_name wafl awa start aggr_name AWA begins collecting workload data for the volumes associated with the specified aggregate. 4. Exit advanced mode: set admin 5. Allow AWA to run until one or more intervals of peak load have occurred. Managing aggregates | 175
AWA analyzes data for up to one rolling week in duration. Running AWA for more than one week will report only on data collected from the previous week. Cache size estimates are based on the highest loads seen during the data collection period. You do not need to ensure that the load is high for the entire data collection period. AWA collects workload statistics for the volumes associated with the specified aggregate. 6. Enter advanced mode: set advanced 7. Display the workload analysis: system node run -node node_name wafl awa print AWA displays the workload statistics and optimal Flash Pool cache size. 8. Stop AWA: system node run -node node_name wafl awa stop All workload data is flushed and is no longer available for analysis. 9. Exit advanced mode: set admin Example In the following example, AWA was run on aggregate “aggr1”. Here is the output of the awa print command after AWA had been running for about 3 days (442 10-minute intervals): ### FP AWA Stats ### Basic Information Aggregate aggr1 Current-time Mon Jul 28 16:02:21 CEST 2014 Start-time Thu Jul 31 12:07:07 CEST 2014 Total runtime (sec) 264682 Interval length (sec) 600 Total intervals 442 In-core Intervals 1024 Summary of the past 442 intervals max Read Throughput 39.695 MB/s Write Throughput 17.581 MB/s Cacheable Read (%) 92 % Cacheable Write (%) 83 % Max Projected Cache Size 114 GiB Projected Read Offload 82 % Projected Write Offload 82 % 176 | Physical Storage Management Guide
Summary Cache Hit Rate vs. Cache Size Size 20% 40% 60% 80% 100% Read Hit 34 51 66 75 82 Write Hit 35 44 53 62 82 The entire results and output of Automated Workload Analyzer (AWA) are estimates. The format, syntax, CLI, results and output of AWA may change in future Data ONTAP releases. AWA reports the projected cache size in capacity. It does not make recommendations regarding the number of data SSDs required. Please follow the guidelines for configuring and deploying Flash Pool; that are provided in tools and collateral documents. These include verifying the platform cache size maximums and minimum number and maximum number of data SSDs. ### FP AWA Stats End ### ________________________________________ The results provide the following pieces of information: • Read Throughput and Write Throughput The throughput measurements can help you identify an aggregate that is receiving a higher amount of traffic. Note that these numbers do not indicate whether that aggregate is I/O bound. • Max Projected Cache Size The size at which the SSD cache would hold every eligible data block that was requested from disk during the AWA run. Note that this does not guarantee a hit for all future I/O operations, because they might request data that is not in the cache. However, if the workload during the AWA run was a typical one, and if your budget allows for it, this would be an ideal size for your Flash Pool cache. • Projected Read Offload and Projected Write Offload The approximate percentages of read and write operations that would have been handled by a Flash Pool cache of the optimal size rather than going to disk (projected cache hit rate). Note that this number is related to the performance increase you would see by converting the aggregate to a Flash Pool aggregate, but not an exact prediction. • Summary Cache Hit Rate vs. Cache Size This table can help you predict the performance impact of decreasing the size of the SSD cache from Max Projected Cache Size. These values are highly impacted by your workload. Depending on whether data that was aged out of the cache was ever accessed again, the impact of decreasing the size of the cache might be large or almost nonexistent. Managing aggregates | 177
You can use this table to find the “sweet spot” of cost versus performance for your workload and budget. Related tasks Determining and enabling volume write-caching eligibility for Flash Pool aggregates on page 178 Related information NetApp Technical Report 4070: Flash Pool Design and Implementation Guide Determining and enabling volume write-caching eligibility for Flash Pool aggregates Understanding whether the FlexVol volumes associated with an aggregate are eligible for write caching can help you ensure that the volumes with high performance requirements can get the maximum performance improvement from having their associated aggregate converted to a Flash Pool aggregate. About this task Flash Pool aggregates employ two types of caching: read caching and write caching. Read caching is available for all volumes. Write caching is available for most volumes, but might be disabled for some volumes due to an internal ID collision. You determine write caching eligibility to help you decide which aggregates are good candidates to become Flash Pool aggregates. You do not need any SSDs to complete this procedure. Steps 1. Attempt to enable the Flash Pool capability on the aggregate: storage aggregate modify aggr_name -hybrid-enabled true 2. Take the applicable action based on the result of Step 1: If... Then... The Flash Pool capability is successfully enabled Disable the Flash Pool capability again: storage aggregate modify aggr_name -hybrid- enabled false You have completed this task. All of the volumes in the aggregate are eligible for write caching. 178 | Physical Storage Management Guide
If... Then... Data ONTAP displays an error message telling you that the aggregate cannot be converted to a Flash Pool aggregate Determine which volumes are not eligible: volume show -volume * -fields hybrid-cache-write- caching-ineligibility-reason -aggregate aggr_name Each volume in the aggregate is listed, along with its reason for ineligibility if it is ineligible. Eligible volumes display a hyphen (“-”). 3. Your next steps depend on your requirements for the ineligible volumes: If you... Then... Do not need write caching enabled on the ineligible volumes You have completed this task. You must use the -force-hybrid- enabled option when you convert the aggregate to a Flash Pool aggregate. Need write caching enabled on the ineligible volumes You must move (or copy and delete) all but one of each set of volumes with the same conflicting ID to another aggregate and then move them back until no more volumes show an ID conflict. Example with ID collisions The following example shows the system output when there are ID collisions: clus1::> vol show -volume * -fields hybrid-cache-write-caching- ineligibility-reason -aggregate aggr1 (volume show) vserver volume hybrid-cache-write-caching-ineligibility-reason ------- -------- ----------------------------------------------- vs0 root_vs0 - vs0 vol1 - vs0 vol2 "ID Collision(27216)" vs0 vol3 "ID Collision(27216)" 4 entries were displayed. Related tasks Determining Flash Pool candidacy and optimal cache size on page 175 Managing aggregates | 179
Changing the RAID type of RAID groups in a Flash Pool aggregate The SSD cache of a Flash Pool aggregate can have a different RAID type than the HDD RAID groups if you need to reduce the parity overhead for your SSD RAID groups. Before you begin • You should understand the considerations for RAID type and spare management for Flash Pool cache. • If you are changing the RAID type of the SSD cache from RAID-DP to RAID4, you should ensure that doing so will not cause you to exceed your cache size limit. About this task You can change the RAID type of the SSD cache for a Flash Pool aggregate using SSD storage pools only when you add the first storage pool allocation units to the aggregate. If the SSD cache has a different RAID type than the HDD RAID groups, the Flash Pool aggregate is considered to have a mixed RAID type, displayed as mixed_raid_type for the aggregate. In this case, the RAID type is also displayed for each RAID group. All HDD RAID groups in a Flash Pool aggregate must have the same RAID type, which should be RAID-DP. Steps 1. Change the RAID type of the SSD cache or HDD RAID groups of the Flash Pool aggregate: storage aggregate modify -aggregate aggr_name -raidtype raid_type - disktype disk_type To change the RAID type of the SSD cache, use -disktype SSD. To change the RAID type of the HDD RAID groups, specify any disk type included in the HDD RAID groups. 2. Verify the RAID groups in your Flash Pool aggregate: storage aggregate show -aggregate aggr_name You also can use the storage aggregate show-status command to obtain more details about the RAID types of the HDD RAID groups and SSD cache of the Flash Pool aggregate. Example In this example, the HDD RAID groups and SSD cache of a Flash Pool aggregate using physical SSDs named “test” initially have a RAID type of RAID-DP. The following command 180 | Physical Storage Management Guide
changes the RAID type of the SSD cache to RAID4, and converts the Flash Pool aggregate to the mixed RAID type: storage aggregate modify -aggregate test -raidtype raid4 -disktype SSD The output from the storage aggregate show-status command shows that the aggregate has a mixed RAID type, the HDD RAID groups have a RAID type of RAID-DP, and the SSD cache has a RAID type of RAID4. storage aggregate show-status test Aggregate test (online, mixed_raid_type, hybrid) (block checksums) Plex /test/plex0 (online, normal, active, pool0) RAID Group /test/plex0/rg0 (normal, block checksums, raid-dp) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ---------- dparity 1.2.3 0 BSAS 7200 827.7GB 828.0GB (normal) parity 1.2.4 0 BSAS 7200 827.7GB 828.0GB (normal) data 1.2.5 0 BSAS 7200 827.7GB 828.0GB (normal) data 1.2.6 0 BSAS 7200 827.7GB 828.0GB (normal) data 1.2.8 0 BSAS 7200 827.7GB 828.0GB (normal) RAID Group /test/plex0/rg1 (normal, block checksums, raid4) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ---------- parity 1.3.3 0 SSD - 82.59GB 82.81GB (normal) data 1.4.0 0 SSD - 82.59GB 82.81GB (normal) data 1.4.1 0 SSD - 82.59GB 82.81GB (normal) data 1.4.2 0 SSD - 82.59GB 82.81GB (normal) Related concepts Considerations for RAID type and spare management for Flash Pool cache on page 132 Related information NetApp Hardware Universe Determining drive and RAID group information for an aggregate Some aggregate administration tasks require that you know what types of drives compose the aggregate, their size, checksum, and status, whether they are shared with other aggregates, and the size and composition of the RAID groups. Step 1. Show the drives for the aggregate, by RAID group: Managing aggregates | 181
storage aggregate show-status aggr_name The drives are displayed for each RAID group in the aggregate. You can see the RAID type of the drive (data, parity, dparity) in the Position column. If the Position column displays shared, then the drive is shared: if it is an HDD, it is a partitioned disk; if it is an SSD, it is part of a storage pool. Example: a Flash Pool aggregate using an SSD storage pool and data partitions cluster1::> storage aggregate show-status nodeA_fp_1 Owner Node: cluster1-a Aggregate: nodeA_fp_1 (online, mixed_raid_type, hybrid) (block checksums) Plex: /nodeA_fp_1/plex0 (online, normal, active, pool0) RAID Group /nodeA_fp_1/plex0/rg0 (normal, block checksums, raid_dp) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ------- shared 2.0.1 0 SAS 10000 472.9GB 547.1GB (normal) shared 2.0.3 0 SAS 10000 472.9GB 547.1GB (normal) shared 2.0.5 0 SAS 10000 472.9GB 547.1GB (normal) shared 2.0.7 0 SAS 10000 472.9GB 547.1GB (normal) shared 2.0.9 0 SAS 10000 472.9GB 547.1GB (normal) shared 2.0.11 0 SAS 10000 472.9GB 547.1GB (normal) RAID Group /nodeA_flashpool_1/plex0/rg1 (normal, block checksums, raid4) (Storage Pool: SmallSP) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ------- shared 2.0.13 0 SSD - 186.2GB 745.2GB (normal) shared 2.0.12 0 SSD - 186.2GB 745.2GB (normal) 8 entries were displayed. Related concepts Understanding root-data partitioning on page 30 How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates on page 137 Relocating aggregate ownership within an HA pair You can change the ownership of aggregates among the nodes in an HA pair without interrupting service from the aggregates. Both nodes in an HA pair are physically connected to each other's disks or array LUNs. Each disk or array LUN is owned by one of the nodes. While ownership of disks temporarily changes when a takeover occurs, the aggregate relocation operations either permanently (for example, if done for load balancing) or temporarily (for example, if done as part of takeover) change the ownership of all disks or array LUNs within an aggregate from one node to the other. The ownership changes without any data-copy processes or physical movement of the disks or array LUNs. 182 | Physical Storage Management Guide
How aggregate relocation works Aggregate relocation takes advantage of the HA configuration to move the ownership of storage aggregates within the HA pair. Aggregate relocation enables storage management flexibility not only by optimizing performance during failover events, but also facilitating system operational and maintenance capabilities that previously required controller failover. Aggregate relocation occurs automatically during manually initiated takeovers to reduce downtime during planned failover events such as nondisruptive software upgrades. You can manually initiate aggregate relocation independent of failover for performance load balancing, system maintenance, and nondisruptive controller upgrades. However, you cannot use the aggregate relocation operation to move ownership of the root aggregate. The following illustration shows the relocation of the ownership of aggregate aggr_1 from Node1 to Node2 in the HA pair: Node2Node1 Aggregate aggr_1 8 disks on shelf sas_1 (shaded grey) Owned by Node1 before relocation Owned by Node2 after relocation The aggregate relocation operation can relocate the ownership of one or more SFO aggregates if the destination node can support the number of volumes in the aggregates. There is only a brief interruption of access to each aggregate. Ownership information is changed one by one for the aggregates. During takeover, aggregate relocation happens automatically after you manually initiate takeover. Before the target controller is taken over, ownership of each of the controller's aggregates is moved, one at a time, to the partner controller. When giveback is initiated, ownership is automatically moved back to the original node. The ‑bypass‑optimization parameter can be used with the storage failover takeover command to suppress aggregate relocation during the takeover. Managing aggregates | 183
Aggregate relocation and Infinite Volumes with SnapDiff enabled The aggregate relocation requires additional steps if the aggregate is currently used by an Infinite Volume with SnapDiff enabled. You must ensure that the destination node has a namespace mirror constituent, and make decisions about relocating aggregates that include namespace constituents. Clustered Data ONTAP 8.3 Infinite Volumes Management Guide Related tasks Relocating ownership of aggregates used by Infinite Volumes on page 156 How root-data partitioning affects aggregate relocation If you have a platform model that uses root-data partitioning, also called shared disks, aggregate relocation processing occurs just as with physical (nonshared) disks. The container disk ownership changes to the destination node during aggregate relocation only if the operation transfers ownership of all partitions on that physical disk to the destination node. This ownership change occurs only with permanent aggregate relocation operations. Ownership changes that occur during negotiated storage failover takeover or giveback events are temporary. Relocating aggregate ownership You can change the ownership of an aggregate only between the nodes within an HA pair. About this task • Because volume count limits are validated programmatically during aggregate relocation operations, it is not necessary to check for this manually. If the volume count exceeds the supported limit, the aggregate relocation operation fails with a relevant error message. • You should not initiate aggregate relocation when system-level operations are in progress on either the source or the destination node; likewise, you should not start these operations during the aggregate relocation. These operations can include the following: ◦ Takeover ◦ Giveback ◦ Shutdown ◦ Another aggregate relocation operation ◦ Disk ownership changes 184 | Physical Storage Management Guide
◦ Aggregate or volume configuration operations ◦ Storage controller replacement ◦ Data ONTAP upgrade ◦ Data ONTAP revert • If you have a MetroCluster configuration, you should not initiate aggregate relocation while disaster recovery operations (switchover, healing, or switchback) are in progress. • If you have a MetroCluster configuration and initiate aggregate relocation on a switched-over aggregate, the operation might fail because it exceeds the DR partner's volume limit count. • You should not initiate aggregate relocation on aggregates that are corrupt or undergoing maintenance. • For All-Flash Optimized FAS80xx-series systems, both nodes in the HA pair must have the All- Flash Optimized personality enabled. Because the All-Flash Optimized configuration supports only SSDs, if one node in the HA pair has HDDs or array LUNs (and therefore, is not configured with the All-Flash Optimized personality), you cannot perform an aggregate relocation from that node to the node with the All- Flash Optimized personality enabled. • If the source node is used by an Infinite Volume with SnapDiff enabled, you must perform additional steps before initiating the aggregate relocation and then perform the relocation in a specific manner. You must ensure that the destination node has a namespace mirror constituent and make decisions about relocating aggregates that include namespace constituents. Clustered Data ONTAP 8.3 Infinite Volumes Management Guide • Before initiating the aggregate relocation, you should save any core dumps on the source and destination nodes. Steps 1. View the aggregates on the node to confirm which aggregates to move and ensure they are online and in good condition: storage aggregate show -node source-node Example The following command shows six aggregates on the four nodes in the cluster. All aggregates are online. Node1 and Node3 form an HA pair and Node2 and Node4 form an HA pair. cluster::> storage aggregate show Aggregate Size Available Used% State #Vols Nodes RAID Status --------- -------- --------- ----- ------- ------ ------ ----------- aggr_0 239.0GB 11.13GB 95% online 1 node1 raid_dp, Managing aggregates | 185
normal aggr_1 239.0GB 11.13GB 95% online 1 node1 raid_dp, normal aggr_2 239.0GB 11.13GB 95% online 1 node2 raid_dp, normal aggr_3 239.0GB 11.13GB 95% online 1 node2 raid_dp, normal aggr_4 239.0GB 238.9GB 0% online 5 node3 raid_dp, normal aggr_5 239.0GB 239.0GB 0% online 4 node4 raid_dp, normal 6 entries were displayed. 2. Issue the command to start the aggregate relocation: storage aggregate relocation start -aggregate-list aggregate-1, aggregate-2... -node source-node -destination destination-node The following command moves the aggregates aggr_1 and aggr_2 from Node1 to Node3. Node3 is Node1's HA partner. The aggregates can be moved only within the HA pair. cluster::> storage aggregate relocation start -aggregate-list aggr_1, aggr_2 -node node1 -destination node3 Run the storage aggregate relocation show command to check relocation status. node1::storage aggregate> 3. Monitor the progress of the aggregate relocation with the storage aggregate relocation show command: storage aggregate relocation show -node source-node Example The following command shows the progress of the aggregates that are being moved to Node3: cluster::> storage aggregate relocation show -node node1 Source Aggregate Destination Relocation Status ------ ----------- ------------- ------------------------ node1 aggr_1 node3 In progress, module: wafl aggr_2 node3 Not attempted yet 2 entries were displayed. node1::storage aggregate> When the relocation is complete, the output of this command shows each aggregate with a relocation status of Done. Related tasks Relocating ownership of aggregates used by Infinite Volumes on page 156 186 | Physical Storage Management Guide
Commands for aggregate relocation There are specific Data ONTAP commands for relocating aggregate ownership within an HA pair. If you want to... Use this command... Start the aggregate relocation process. storage aggregate relocation start Monitor the aggregate relocation process storage aggregate relocation show Related information Clustered Data ONTAP 8.3 Commands: Manual Page Reference Key parameters of the storage aggregate relocation start command The storage aggregate relocation start command includes several key parameters used when relocating aggregate ownership within an HA pair. Parameter Meaning -node nodename Specifies the name of the node that currently owns the aggregate -destination nodename Specifies the destination node where aggregates are to be relocated -aggregate-list aggregate name Specifies the list of aggregate names to be relocated from source node to destination node (This parameter accepts wildcards) -override-vetoes true|false Specifies whether to override any veto checks during the relocation operation Use of this option can potentially lead to longer client outage, or aggregates and volumes not coming online after the operation. Managing aggregates | 187
Parameter Meaning -relocate-to-higher-version true| false Specifies whether the aggregates are to be relocated to a node that is running a higher version of Data ONTAP than the source node • You cannot perform an aggregate relocation from a node running Data ONTAP 8.2 to a node running Data ONTAP 8.3 or higher using the ‑relocate‑to‑higher-version true parameter. You must first upgrade the source node to Data ONTAP 8.2.1 or higher before you can perform an aggregate relocation operation using this parameter. Similarly, you must upgrade to Data ONTAP 8.2.1 or higher before you can upgrade to Data ONTAP 8.3. • Although you can perform aggregate relocation between nodes running different minor versions of Data ONTAP (for example, 8.2.1 to 8.2.2, or 8.2.2 to 8.2.1), you cannot perform an aggregate relocation operation from a higher major version to a lower major version (for example, 8.3 to 8.2.2). -override-destination-checks true| false Specifies if the aggregate relocation operation should override the check performed on the destination node Use of this option can potentially lead to longer client outage, or aggregates and volumes not coming online after the operation. Veto and destination checks during aggregate relocation In aggregate relocation operations, Data ONTAP determines whether aggregate relocation can be completed safely. If aggregate relocation is vetoed, you must check the EMS messages to determine the cause. Depending on the reason or reasons, you can decide whether you can safely override the vetoes. The storage aggregate relocation show command displays the aggregate relocation progress and shows which subsystem, if any, vetoed the relocation. Soft vetoes can be overridden, but hard vetoes cannot be, even if forced. You can review the EMS details for any giveback vetoes by using the following command: 188 | Physical Storage Management Guide
event log show -node * -event gb* You can review the EMS details for aggregate relocation by using the following command: event log show -node * -event arl* The following tables summarize the soft and hard vetoes, along with recommended workarounds: Veto checks during aggregate relocation Vetoing subsystem module Workaround Vol Move Relocation of an aggregate is vetoed if any volumes hosted by the aggregate are participating in a volume move that has entered the cutover state. Wait for the volume move to complete. If this veto is overridden, cutover resumes automatically once the aggregate relocation completes. If aggregate relocation causes the move operation to exceed the number of retries (the default is 3), then the user needs to manually initiate cutover using the volume move trigger-cutover command. Backup Relocation of an aggregate is vetoed if a dump or restore job is in progress on a volume hosted by the aggregate. Wait until the dump or restore operation in progress is complete. If this veto is overridden, the backup or restore operation is aborted and must be restarted by the backup application. Lock manager To resolve the issue, gracefully shut down the CIFS applications that have open files, or move those volumes to a different aggregate. Overriding this veto results in loss of CIFS lock state, causing disruption and data loss. Lock Manager NDO Wait until the locks are mirrored. This veto cannot be overridden; doing so disrupts Microsoft Hyper-V virtual machines. RAID Check the EMS messages to determine the cause of the veto: If disk add or disk ownership reassignment operations are in progress, wait until they complete. If the veto is due to a mirror resync, a mirror verify, or offline disks, the veto can be overridden and the operation restarts after giveback. Managing aggregates | 189
Destination checks during aggregate relocation Vetoing subsystem module Workaround Disk Inventory Relocation of an aggregate fails if the destination node is unable to see one or more disks belonging to the aggregate. Check storage for loose cables and verify that the destination can access disks belonging to the aggregate being relocated. This check cannot be overridden. WAFL Relocation of an aggregate fails if the relocation would cause the destination to exceed its limits for maximum volume count or maximum volume size. This check cannot be overridden. Lock Manager NDO Relocation of an aggregate fails if: • The destination does not have sufficient lock manager resources to reconstruct locks for the relocating aggregate. • The destination node is reconstructing locks. Retry aggregate relocation after a few minutes. This check cannot be overridden. Lock Manager Permanent relocation of an aggregate fails if the destination does not have sufficient lock manager resources to reconstruct locks for the relocating aggregate. Retry aggregate relocation after a few minutes. This check cannot be overridden. RAID Check the EMS messages to determine the cause of the failure: • If the failure is due to an aggregate name or UUID conflict, troubleshoot and resolve the issue. This check cannot be overridden. Relocating an aggregate fails if the relocation would cause the destination to exceed its limits for maximum aggregate count, system capacity, or aggregate capacity. You should avoid overriding this check. Assigning aggregates to SVMs If you assign one or more aggregates to a Storage Virtual Machine (SVM, formerly known as Vserver), then you can use only those aggregates to contain volumes for that SVM. Assigning 190 | Physical Storage Management Guide
aggregates to your SVMs is particularly important in a multi-tenancy environment or when you use Infinite Volumes. Before you begin The SVM and the aggregates you want to assign to that SVM must already exist. About this task Assigning aggregates to your SVMs helps you keep your SVMs isolated from each other; this is especially important in a multi-tenancy environment. If you use Infinite Volumes, or plan to use them in the future, you must assign aggregates to your SVMs to keep your Infinite Volumes from impacting each other and any FlexVol volumes in your cluster. Steps 1. Check the list of aggregates already assigned to the SVM: vserver show -fields aggr-list The aggregates currently assigned to the SVM are displayed. If there are no aggregates assigned, “-” is displayed. 2. Add or remove assigned aggregates, depending on your requirements: If you want to... Use this command... Assign additional aggregates vserver add-aggregates Unassign aggregates vserver remove-aggregates The listed aggregates are assigned to or removed from the SVM. If the SVM already has volumes that use an aggregate that is not assigned to the SVM, a warning message is displayed, but the command is completed successfully. Any aggregates that were already assigned to the SVM and that were not named in the command are unaffected. Example In the following example, the aggregates aggr1 and aggr2 are assigned to SVM svm1: vserver add-aggregates -vserver svm1 -aggregates aggr1,aggr2 Methods to create space in an aggregate If an aggregate runs out of free space, various problems can result that range from loss of data to disabling a volume's guarantee. There are multiple ways to make more space in an aggregate. All of the methods have various consequences. Prior to taking any action, you should read the relevant section in the documentation. Managing aggregates | 191
The following are some common ways to make space in an aggregate, in order of least to most consequences: • Add disks to the aggregate. • Move some volumes to another aggregate with available space. • Shrink the size of volume-guaranteed volumes in the aggregate. You can do this manually or with the autoshrink option of the autosize capability. • Change volume guarantee types to none on volumes that are using large amounts of space (large volume-guaranteed volumes with large reserved files) so that the volumes take up less space in the aggregate. A volume with a guarantee type of none has a smaller footprint in the aggregate than a volume with a guarantee type of volume. The Volume Guarantee row of the volume show- footprint command output shows whether a volume is reserving a large amount of space in the aggregate due to its guarantee. • Delete unneeded volume Snapshot copies if the volume's guarantee type is none. • Delete unneeded volumes. • Enable space-saving features, such as deduplication or compression. • (Temporarily) disable features that are using a large amount of metadata (visible with the volume show-footprint command). Determining which volumes reside on an aggregate You might need to determine which FlexVol volumes or Infinite Volume constituents reside on an aggregate before performing operations on the aggregate, such as relocating it or taking it offline. About this task Infinite Volume constituents are somewhat similar to FlexVol volumes, but you usually do not manage them directly. For more information about Infinite Volumes and constituents, see the Clustered Data ONTAP Infinite Volumes Management Guide. Step 1. Enter the appropriate command, depending on whether your system has Infinite Volumes: If your system... Then use this command... Does not have Infinite Volumes volume show -aggregate aggregate_name 192 | Physical Storage Management Guide
If your system... Then use this command... Has Infinite Volumes volume show -is-constituent * -aggregate aggregate_name All volumes (and, if you have Infinite Volumes, constituents) that reside on the specified aggregate are displayed. Determining whether a Flash Pool aggregate is using an SSD storage pool You manage Flash Pool aggregates differently when they use SSD storage pools to provide their cache than when they use discrete SSDs. Step 1. Display the aggregate's drives by RAID group: storage aggregate show-status aggr_name If the aggregate is using one or more SSD storage pools, the value for the Position column for the SSD RAID groups is displayed as Shared, and the name of the storage pool is displayed next to the RAID group name. Commands for managing aggregates You use the storage aggregate command to manage your aggregates. If you want to... Use this command... Display the size of the cache for all Flash Pool aggregates storage aggregate show -fields hybrid-cache-size-total -hybrid- cache-size-total >0 Display disk information and status for an aggregate storage aggregate show-status Display spare disks by node storage aggregate show-spare-disks Display the root aggregates in the cluster storage aggregate show -has-mroot true Display basic information and status for aggregates storage aggregate show Bring an aggregate online storage aggregate online Managing aggregates | 193
If you want to... Use this command... Delete an aggregate storage aggregate delete Put an aggregate into the restricted state storage aggregate restrict Rename an aggregate storage aggregate rename Take an aggregate offline storage aggregate offline Change the RAID type for an aggregate storage aggregate modify -raidtype Related information Clustered Data ONTAP 8.3 Commands: Manual Page Reference 194 | Physical Storage Management Guide
Storage limits There are limits for storage objects that you should consider when planning and managing your storage architecture. Limits are listed in the following sections: • Aggregate limits • RAID group limits Aggregate limits Storage object Limit Native storage Storage arrays Virtual storage (Data ONTAP-v) Aggregates Maximum per node1 100 100 60 Maximum size2 Model- dependent Model- dependent 16 TB Minimum size3 RAID-DP: 5 disks RAID4: 3 disks Model- dependent 1 disk Aggregates (mirrored) Maximum suggested per node4 64 64 N/A RAID groups Maximum per aggregate 150 150 60 Notes: 1. In an HA configuration, this limit applies to each node individually, so the overall limit for the pair is doubled. 2. See the Hardware Universe. 3. For root aggregates using physical disks, the minimum size is 3 disks for RAID-DP and 2 disks for RAID4. See the Hardware Universe for the minimum aggregate size for storage arrays. 195
4. You can create more than 64 mirrored aggregates on a node, but doing so could cause plex synchronization problems after certain types of failures. RAID group limits For maximum and default RAID group sizes, see the Hardware Universe. Limit Native storage Storage arrays Virtual storage (Data ONTAP-v) Maximum per node 400 400 60 Maximum per aggregate 150 150 60 196 | Physical Storage Management Guide
Copyright information Copyright © 1994–2015 NetApp, Inc. All rights reserved. Printed in the U.S. No part of this document covered by copyright may be reproduced in any form or by any means— graphic, electronic, or mechanical, including photocopying, recording, taping, or storage in an electronic retrieval system—without prior written permission of the copyright owner. Software derived from copyrighted NetApp material is subject to the following license and disclaimer: THIS SOFTWARE IS PROVIDED BY NETAPP "AS IS" AND WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, WHICH ARE HEREBY DISCLAIMED. IN NO EVENT SHALL NETAPP BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. NetApp reserves the right to change any products described herein at any time, and without notice. NetApp assumes no responsibility or liability arising from the use of products described herein, except as expressly agreed to in writing by NetApp. The use or purchase of this product does not convey a license under any patent rights, trademark rights, or any other intellectual property rights of NetApp. The product described in this manual may be protected by one or more U.S. patents, foreign patents, or pending applications. RESTRICTED RIGHTS LEGEND: Use, duplication, or disclosure by the government is subject to restrictions as set forth in subparagraph (c)(1)(ii) of the Rights in Technical Data and Computer Software clause at DFARS 252.277-7103 (October 1988) and FAR 52-227-19 (June 1987). 197
Trademark information NetApp, the NetApp logo, Go Further, Faster, ASUP, AutoSupport, Campaign Express, Cloud ONTAP, clustered Data ONTAP, Customer Fitness, Data ONTAP, DataMotion, Fitness, Flash Accel, Flash Cache, Flash Pool, FlashRay, FlexArray, FlexCache, FlexClone, FlexPod, FlexScale, FlexShare, FlexVol, FPolicy, GetSuccessful, LockVault, Manage ONTAP, Mars, MetroCluster, MultiStore, NetApp Insight, OnCommand, ONTAP, ONTAPI, RAID DP, SANtricity, SecureShare, Simplicity, Simulate ONTAP, Snap Creator, SnapCopy, SnapDrive, SnapIntegrator, SnapLock, SnapManager, SnapMirror, SnapMover, SnapProtect, SnapRestore, Snapshot, SnapValidator, SnapVault, StorageGRID, Tech OnTap, Unbound Cloud, and WAFL are trademarks or registered trademarks of NetApp, Inc., in the United States, and/or other countries. A current list of NetApp trademarks is available on the web at http://www.netapp.com/us/legal/netapptmlist.aspx. Cisco and the Cisco logo are trademarks of Cisco in the U.S. and other countries. All other brands or products are trademarks or registered trademarks of their respective holders and should be treated as such. 198 | Physical Storage Management Guide
How to send comments about documentation and receive update notification You can help us to improve the quality of our documentation by sending us your feedback. You can receive automatic notification when production-level (GA/FCS) documentation is initially released or important changes are made to existing production-level documents. If you have suggestions for improving this document, send us your comments by email to doccomments@netapp.com. To help us direct your comments to the correct division, include in the subject line the product name, version, and operating system. If you want to be notified automatically when production-level documentation is released or important changes are made to existing production-level documents, follow Twitter account @NetAppDoc. You can also contact us in the following ways: • NetApp, Inc., 495 East Java Drive, Sunnyvale, CA 94089 U.S. • Telephone: +1 (408) 822-6000 • Fax: +1 (408) 822-4501 • Support telephone: +1 (888) 463-8277 199
Index A ACP how you use to increase storage availability for SAS- connected disk shelves 24 active-active configurations setting up with root-data partitioning 39 adding key management servers 90 aggregate ownership relocation of 182 aggregate relocation benefits of 183 commands for 187 effect on root-data partitioning 184 effect on shared disks 184 how it works 183 Infinite Volumes 156 monitoring progress of 188 overriding a veto of 188 aggregate show-space command how to determine aggregate space usage by using 148 aggregates adding physical drives or array LUNs to 163 assigning to SVMs 190 associated with Infinite Volumes 154 changing size of RAID groups for 115 commands for displaying space usage information 53 commands for managing 193 configuration requirements for multi-disk carrier shelves 29 considerations for sizing RAID groups 114 considerations for using disks from multi-disk carriers in 29 creating Flash Pool 171 creating Flash Pool using SSD storage pools 173 creating SSD storage pool for Flash Pool 140 creating using physical drives 159 creating using root-data partitioning 161 creating using shared HDDs 161 description and characteristics of 127 determination of checksum type of array LUN 148 determining candidacy and optimal cache size for Flash Pool 175 determining drive information for 181 determining RAID group information for 181 determining which volumes reside on 192 disk speeds supported by Data ONTAP 13 effect of SVM on selection 144 Flash Pool, defined 130 Flash Pool, determining volume write-caching eligibility for 178 Flash Pool, determining whether using a storage pool 193 Flash Pool, how they work 130 how associated with SVMs with Infinite Volume 154 how cache capacity is calculated for Flash Pool 133 how drive checksum types affect management 14 how root-data partitioning affects storage management 31 how storage classes use for Infinite Volumes 153 how storage pools increase cache allocation flexibility for Flash Pool 137 how to determine space usage in 148 how you use storage pools with Flash Pool 139 increasing size of, when they use shared HDDs 165 increasing the size of with physical drives 163 introduction to managing 159 maximum and minimum size of 195 maximum per node 195 maximum size, methods of calculating 13 methods of creating space in 191 mirrored, explained 129 mirrored, maximum per node 195 ownership change 184 relocating ownership of 156 relocation of 183, 184 replacing disks used in 43 requirements and best practices for using storage pools with Flash Pool 139 requirements for Infinite Volumes 152 requirements for storage classes 153 requirements for using Flash Pool 131 root, which drives are partitioned for use in 32 rules about mixing storage types in 147 rules for mixing HDD types in 146 rules for storage array families 147 sharing between FlexVol volumes and Infinite Volumes 152 tips for creating and backing up, for sensitive data 21 200 | Physical Storage Management Guide
unmirrored, explained 127 ways to use disks with mixed speeds in 145 what happens when adding storage to 170 All-Flash Optimized personality how it affects node behavior 25 Alternate Control Path See ACP array LUN ownership how it works 66 array LUNs adding to aggregates 163 considerations for sizing RAID groups for 114 how to control selection from heterogeneous storage 145 reasons to assign ownership of 54, 66 systems running Data ONTAP that can use 64 verifying back-end configuration for 72 assigning aggregates 154 ATA drives how Data ONTAP reports disk types 10 authentication keys changing 96 deleting 98 how Storage Encryption uses 82 removing management server that stores 93 retrieving 97 autoassignment See automatic ownership assignment Automated Workflow Analyzer See AWA automatic ownership assignment configuring for disks 61 guidelines for disks 58 how it works for disks 56 when it is invoked 57 which policy to use for disk 56 automatic RAID-level scrubs changing the schedule for 122 availability how Data ONTAP uses RAID to ensure data 108 AWA determining Flash Pool optimal cache size by using 175 AZCS type checksums configuration rules 14 effect on aggregate management 14 effect on spare management 14 B back-end configurations verifying 72 BCS type checksums configuration rules 14 effect on aggregate management 14 effect on spare management 14 benefits of Storage Encryption 83 best practices for using storage pools 139 block checksum type changing for array LUNs 76 BSAS drives how Data ONTAP reports disk types 10 C cache capacity how calculated for Flash Pool aggregates 133 cache size determining impact to, when adding SSDs to storage pool 143 determining optimal for Flash Pool aggregates 175 cache storage requirements and best practices for using storage pools for Flash Pool aggregate 139 caches comparison of Flash Pool and Flash Cache 133 caching policies modifying 136 using Flash Pool aggregates 135 capacity how it is allocated in new Infinite Volumes 155 methods of calculating aggregate and system 13 carriers determining when to remove multi-disk 28 how Data ONTAP avoids RAID impact when removing multi-disk 27 spare requirements for multi-disk 28 certificates installing replacement SSL 99 installing SSL, on storage systems 87 preventing SSL expiration issues 98 removing old SSL 99 SSL requirements 87 changing authentication keys 96 RAID group size 115 Index | 201
RAID type for Flash Pool cache 180 checksums checking the type 76 configuration rules 14 rules for aggregates 148 type, changing for array LUNs 76 type, effect on aggregate and spare management 14 commands aggregate management, list of 193 disk management, list of 51 for displaying aggregate space usage information 53 for displaying FlexVol volume space usage information 53 for displaying information about storage shelves 52 how you use wildcard character with disk ownership 62 SSD storage pool management, list of 143 storage aggregate 51, 53, 193 storage disk 51 storage pool 143 storage pool management, list of 143 storage shelves, displaying information about 52 volume show-footprint 53 volume show-space 53 volume snapshot 53 comments how to send feedback about documentation 199 composition changing array LUN 78 configuring automatic ownership assignment of disks 61 connection types how disks can be combined for SAS 12 supported storage 12 constituents determining which ones reside on an aggregate 192 continuous media scrubbing how Data ONTAP uses, to prevent media errors 24 impact on system performance 24 reasons it should not replace scheduled RAID-level disk scrubs 24 core files spare disk requirement for 117 creating aggregates using physical drives 159 Flash Pool aggregates 171 Flash Pool aggregates using SSD storage pools 173 current owner disk ownership type, defined 54 D data how Data ONTAP uses RAID to protect and ensure availability 108 tips for creating and backing up aggregates containing sensitive 21 using sanitization to remove from disks 48 data at rest introduction to securing with Storage Encryption 81 data disks removing 47 data integrity how RAID-level disk scrubs verify 122 Data ONTAP disk types comparison with industry standard 10 Data ONTAP Edge how disk ownership works 57 Data ONTAP-v how disk ownership works 57 data protection in case of disk loss or theft 83 through emergency shredding 84 when moving disks to end-of-life 84 when returning disks to vendors 83 data reconstruction controlling performance impact of RAID 124 data shredding performing emergency, on disks using Storage Encryption 104 degraded mode 119 deleting authentication keys 98 destroying data on disks using Storage Encryption 102 disk performance monitors 21 disk connection types how disks can be combined for SAS 12 disk operations with SEDs 82 disk ownership application to array LUNs 55, 66, 68 application to disks 55, 66 assigning array LUNs 71 configuring automatic assignment 61 how it works 66 how it works for Data ONTAP Edge 57 how it works for Data ONTAP-v 57 ownership 202 | Physical Storage Management Guide
removing array LUN ownership 79 removing array LUN ownership 79 disk ownership commands how you use wildcard character with 62 Disk Qualification Package when you need to update 42 disk remove -w removing an array LUN 79 disk sanitization introduction to how it works 19 process described 19 when it cannot be performed 20 disk scrubbing reasons continuous media scrubbing should not replace scheduled RAID-level 24 disk shelves aggregate configuration requirements for multi-disk carrier 29 commands for displaying information about 52 configuration requirements for multi-disk carrier 29 how you use ACP to increase storage availability for SAS-connected 24 requirements for using multi-disk carrier 27 disk slicing understanding for entry-level platforms 30 disk types how to control selection from heterogeneous storage 145 disks adding to a node 41 adding to aggregates 163 assigning ownership for 58 commands for managing 51 considerations for removing from storage systems 45 considerations for using, from multi-disk carriers in aggregates 29 data, converting to spare 44 determining information about, for aggregates 181 displaying information about Storage Encryption 95 evacuation process, about 27 guidelines for assigning ownership 58 how automatic ownership assignment works 56 how available for Data ONTAP use 55, 66 how Data ONTAP handles failed, with available hot spares 120 how Data ONTAP handles failed, with no available hot spare 121 how Data ONTAP reduces failures using Rapid RAID Recovery 21 how Data ONTAP reports types 10 how low spare warnings can help you manage spare 120 how RAID-level scrubs verify data integrity 122 how root-data partitioning affects storage management 31 how shared HDDs work 30 how they can be combined for SAS connection type 12 how to control selection from heterogeneous storage 145 introduction to how DATA ONTAP works with heterogeneous storage 145 introduction to managing ownership for 54 loop IDs for FC-AL connected, about 18 managing using Data ONTAP 10 matching spares defined 118 minimum required hot spare 117 performing emergency data shredding on Storage Encryption 104 physical, adding to aggregates 163 RAID drive types, defined 18, 113 RAID protection levels for 108 reasons to assign ownership of 54, 66 removing data 47 removing failed 45 removing hot spares 46 removing ownership from 60 replacing in aggregate 43 replacing self-encrypting 44 requirements for using partitioned 34 rules for mixing HDD types in aggregates 146 sanitization process described 19 sanitization, what happens if interrupted 20 sanitizing 102 setting state to end-of-life 103 slicing, understanding for entry level platforms 30 spare requirements for multi-disk carrier 28 spare, appropriate 118 speeds supported by Data ONTAP 13 SSD and HDD capability differences 26 stopping sanitization 51 types of ownership for 54 using sanitization to remove data from 48 ways to mix speed of, in aggregates 145 what happens when Data ONTAP takes them offline 21 when automatic ownership assignment is invoked 57 when sanitization cannot be performed 20 when they can be put into maintenance center 23 Index | 203
when you need to update the Disk Qualification Package for 42 which autoassignment policy to use for 56 displaying key management server information 92 key management server status 91 Storage Encryption disk information 95 documentation how to receive automatic notification of changes to 199 how to send feedback about 199 DQP See Disk Qualification Package DR home owner disk ownership type, defined 54 drive errors how the maintenance center helps prevent 22 drive types RAID, defined 18, 113 drives considerations for sizing RAID groups for 114 how Data ONTAP handles failed, with available hot spares 120 how Data ONTAP handles failed, with no available hot spares 121 how low spare warnings can help you manage spare 120 name formats 14 pre-cluster name formats 15 rules for mixing types in Flash Pool aggregates 147 which are partitioned for root-data partitioning 32 See also disks E emergency data shredding data protection through 84 performing on disks using Storage Encryption 104 end-of-life setting disk state to 103 entry-level platforms understanding root-data partitioning for 30 errors how Data ONTAP uses media scrubbing to prevent media 24 how the maintenance center helps prevent drive 22 evacuation process for disks, about 27 existing storage pools considerations for adding SSDs to 142 external key management servers defined 81 displaying information about 92 F failed disks removing 45 family defined 147 FC storage connection type how disks can be combined for 12 support for 12 FCAL drives how Data ONTAP reports disk types 10 feedback how to send comments about documentation 199 Fibre Channel See FC Flash Cache compared with Flash Pool aggregates 133 Flash Pool aggregates caching policies 135 changing RAID type 180 compared with Flash Cache 133 creating 171 creating SSD storage pool for 140 creating using SSD storage pools 173 defined 130 determining candidacy and optimal cache size for 175 determining whether using a storage pool 193 how cache capacity is calculated for 133 how storage pools increase cache allocation flexibility for 137 how they work 130 how you use storage pools with 139 RAID type, considerations for 132 requirements and best practices for using storage pools with 139 requirements for using 131 rules for mixing drive types in 147 spare management, considerations for 132 volume write-caching eligibility, determining 178 Flash Pool SSD partitioning how it increases cache allocation flexibility for Flash Pool aggregates 137 FlexVol volumes aggregate sharing with Infinite Volumes 152 204 | Physical Storage Management Guide
commands for displaying space usage information 53 creating aggregates using physical drives 159 determining which ones reside on an aggregate 192 effect of SVM on aggregate selection 144 formats drive name 14 pre-cluster drive name 15 FSAS drives how Data ONTAP reports disk types 10 G groups RAID, how they work 113 guidelines assigning disk ownership 58 H hard disk drives See HDDs HDD RAID groups sizing considerations for 114 HDDs assigning ownership for shared 59 capability differences with SSDs 26 creating aggregates using shared 161 increasing size of aggregates that use shared 165 rules for mixing types in aggregates 146 shared, how they work 30 speeds supported by Data ONTAP 13 standard layouts for shared 32 heterogeneous storage how to control disk selection from 145 introduction to how Data ONTAP works with 145 high-performance aggregates Flash Pool, defined 130 home owner disk ownership type, defined 54 hot spares appropriate 118 defined 117 how Data ONTAP handles failed disks with available 120 how Data ONTAP handles failed disks with no available 121 matching, defined 118 minimum needed 117 removing 46 what disks can be used as 118 hybrid aggregates See Flash Pool aggregates I IDs about loop, for FC-AL connected disks 18 increasing aggregate size using physical drives 163 Infinite Volumes aggregate relocation 156 aggregate requirements 152 associated aggregates 154 capacity allocation 155 creating aggregates using physical drives 159 determining which constituents reside on an aggregate 192 how storage classes use aggregates for 153 how to determine space usage for constituents 150 relocating aggregates 156 space allocation 155 InfiniteVol See Infinite Volumes information how to send feedback about improving documentation 199 initialization performing on node to configure root-data partitioning 34 installing replacement SSL certificates 99 SSL certificates on storage systems 87 K Key Management Interoperability Protocol using for communication with key management servers 81 key management servers adding 90 displaying information about 92 displaying status 91 external, defined 81 introduction to using SSL for secure communication 86 precautions taken against unreachability during boot process 94 removing 93 verifying links 91 Index | 205
keys changing authentication 96 how Storage Encryption uses authentication 82 retrieving authentication 97 KMIP See Key Management Interoperability Protocol L layouts standard shared HDD 32 levels RAID protection, for disks 108 licenses for using array LUNs 65 installing for array LUN use 65 V_StorageAttach 65 limitations Storage Encryption 84 limits aggregate storage 195 FlexClone file and LUN storage 195 RAID group storage and size 195 volume storage 195 loop IDs about FC-AL connected disk 18 loops configuring automatic ownership assignment for 61 low spare warnings how they can help you manage spare drives 120 LUNs (array) assigning ownership of 71 changing checksum type 76 changing ownership assignment 73 changing size or composition 78 checking the checksum type of 76 Data ONTAP owning 68 Data ONTAP RAID groups with 116 examples of when Data ONTAP can use 69 how available for Data ONTAP use 55, 66 managing through Data ONTAP 64 names format of 74 prerequisites to changing composition 77 prerequisites to changing size 77 RAID protection for 109 reasons to assign ownership of 54, 66 reasons you might assign to a system 68 requirements before removing a system running Data ONTAP from service 79 rules about mixing storage types in aggregates 147 setting them up in Data ONTAP 64 systems running Data ONTAP that can use 64 verifying back-end configuration for 72 LUNs, array See LUNs (array) M maintenance center how it helps prevent drive errors 22 when disks go into 23 management servers adding key 90 displaying information about key 92 removing authentication key 93 verifying server links of key 91 managing Storage Encryption 90 manual RAID-level scrubs how to run 123 matching spare disks defined 118 media errors how Data ONTAP uses media scrubbing to prevent 24 media scrubbing how Data ONTAP uses, to prevent media errors 24 impact on system performance 24 reasons it should not replace scheduled RAID-level disk scrubs 24 mirror verification controlling performance impact 126 mirrored aggregates explained 129 modifying caching policies 136 MSATA drives how Data ONTAP reports disk types 10 MSIDs rekeying SEDs to 100 multi-disk carrier shelves aggregate configuration requirements for 29 configuration requirements for 29 in aggregates, considerations for using disks from 29 requirements for using 27 multi-disk carriers determining when to remove 28 how Data ONTAP handles when removing 27 spare requirements for 28 206 | Physical Storage Management Guide
N name format array LUNs 74 names formats for drive 14 formats for pre-cluster drive 15 new storage pools considerations for adding SSDs to 142 NL-SAS drives how Data ONTAP reports disk types 10 node behavior how All-Flash Optimized personality affects 25 nodes adding disks to 41 how to control disk selection from heterogeneous storage on 145 initializing for root-data partitioning 34 O offline what happens when Data ONTAP takes disks 21 original owner disk ownership type, defined 54 owner disk ownership type, defined 54 ownership assigning array LUNs 71 assigning for disks 58 assigning for shared HDDs 59 automatically assigning to a stack or shelf 61 disk, types of 54 guidelines for assigning disk 58 how it works for disks and array LUNs 66 introduction to managing, for disks 54 reasons to assign disk and array LUN 54, 66 removing from disks 60 ownership assignment when it is invoked for disks 57 ownership commands how you use wildcard character with disk 62 P partitioning how Flash Pool SSD increases cache allocation flexibility for 137 root-data, assigning ownership for HDDs partitioned for 59 root-data, creating aggregates using 161 root-data, how it works 30 root-data, how storage management is affected by 31 root-data, increasing the size of an aggregate using 165 root-data, requirements for using 34 root-data, standard layouts for 32 root-data, understanding 30 root-data, which drives are partitioned for 32 setting up active-active configuration on node using root-data 39 partitions correcting misaligned spare 169 performance controlling impact of RAID data reconstruction 124 impact of media scrubbing on system 24 persistent reservations releasing all 79 physical drives using to create aggregates 159 physical secure IDs See PSIDs platforms understanding root-data partitioning for entry-level 30 plex resynchronization controlling performance impact of 125 plexes mirrored aggregate, explained 129 policies autoassignment, which to use for disks 56 pools considerations for when to use SSD storage 138 requirements and best practices for using storage 139 protection how Data ONTAP uses to RAID for data 108 RAID levels for disks 108 PSIDs how the factory resets SEDs with 105 SEDs that include the functionality 106 using to reset SED to factory original settings 106 R RAID avoiding impact to, when replacing multi-disk carriers 27 data reconstruction, controlling performance impact 124 drive types defined 18, 113 Index | 207
how Data ONTAP to protect data and data availability 108 how disk scrubs verify data integrity 122 operations, controlling performance impact 123 protection levels for disks 108 protection with SyncMirror and 110 scrub, controlling performance impact 124 scrubs, changing the schedule for 122 type, changing for Flash Pool cache 180 type, determining for Flash Pool cache 132 RAID groups changing size of 115 definition 113 determining information about, for aggregates 181 how they work 113 maximum per aggregate 195 maximum per node 196 naming convention 114 sizing considerations for 114 what happens when adding storage to aggregates in 170 with array LUNs, considerations 116 RAID protection for array LUNs 109 RAID-DP described 108 RAID-level scrubs automatic schedule, changing 122 how to run manual 123 reasons media scrubbing should not replace scheduled 24 raid.timeout option considerations for changing 121 RAID0 how Data ONTAP uses for array LUNs 109 use by Data ONTAP 109 RAID4 described 109 Rapid RAID Recovery how Data ONTAP reduces disk failures using 21 rekeying SEDs to MSID 100 relocating aggregates Infinite Volumes 156 relocation aggregate ownership 182, 184 of aggregates 182–184 removing data disks 47 data, using disk sanitization 48 failed disks 45 hot spare disks 46 key management servers 93 multi-disk carriers, determining when it is safe 28 old SSL certificates 99 replacing disks in aggregates 43 requirements Flash Pool aggregate use 131 for using root-data partitioning 34 for using storage pools 139 Infinite Volumes, aggregate 152 resetting SEDs to factory original settings 106 resynchronization controlling performance impact of plex 125 retrieving authentication keys 97 root-data partitioning assigning ownership for HDDs using 59 correcting misaligned spare partitions for 169 creating aggregates using 161 effect on aggregate relocation 184 how it works 30 how storage management is affected by 31 increasing the size of an aggregate using 165 initializing node to configure 34 requirements for using 34 setting up active-active configuration on node using 39 standard layouts for 32 understanding 30 which drives are partitioned for 32 rules for mixing drive types in Flash Pool aggregates 147 for mixing HDD types in aggregates 146 S sanitization disk process described 19 disk, introduction to how it works 19 stopping disk 51 tips for creating and backing up aggregates containing sensitive data 21 using to remove data from disks 48 what happens if interrupted 20 when it cannot be performed 20 sanitizing disks 102 208 | Physical Storage Management Guide
SAS storage connection type, support for 12 SAS drives how Data ONTAP reports disk types 10 SAS-connected shelves how disks can be combined for 12 how you use ACP to increase storage availability for 24 SATA drives how Data ONTAP reports disk types 10 scrubbing how Data ONTAP uses media, to prevent media errors 24 impact of media, on system performance 24 media, reasons it should not replace scheduled RAID-level disk scrubs 24 scrubs changing automatic schedule for 122 controlling performance impact of RAID 124 how to run manual RAID-level 123 RAID-level, how they verify data integrity 122 secure communication introduction to using SSL for 86 SEDs disk operations with 82 how Storage Encryption works with 82 how the factory resets with PSID 105 replacing 44 resetting to factory original settings 106 returning to unprotected mode 100 that have PSID functionality 106 self-encrypting disks See SEDs serial-attached SCSI See SAS servers adding key management 90 displaying information about key management 92 removing authentication key management 93 verifying server links of key management 91 setting up array LUNs 64 Storage Encryption 85 setup wizards running the Storage Encryption 88 shared HDDs assigning ownership for 59 creating aggregates using 161 how they work 30 increasing size of aggregates that use 165 understanding 30 shared layouts standard HDD 32 shared SSDs See storage pools shelves aggregate configuration requirements for multi-disk carrier 29 commands for displaying information about 52 configuration requirements for multi-disk carrier 29 configuring automatic ownership assignment for 61 how you use ACP to increase storage availability for SAS-connected 24 requirements for using multi-disk carrier 27 shredding performing emergency data, on disks using Storage Encryption 104 size changing array LUN size 77 sizes changing array LUN 78 sizing RAID groups, considerations for 114 Snapshot reserve commands for displaying size of 53 solid-state disks See SSDs space commands for displaying usage information 53 how it is allocated in new Infinite Volumes 155 methods of creating in an aggregate 191 space usage how to determine and control volume, in aggregates 150 how to determine in an aggregate 148 spare array LUNs changing array LUN assignment 73 changing ownership assignment 73 checking the type 76 disk ownership 73 spare disks appropriate 118 defined 117 how checksum types affect management 14 how Data ONTAP handles failed disks with available 120 how Data ONTAP handles failed disks with no available 121 how low spare warnings can help you manage 120 managing for Flash Pool cache 132 Index | 209
matching, defined 118 minimum needed 117 removing 46 removing ownership from 60 requirements for multi-disk carriers 28 what disks can be used as 118 spare partitions correcting misaligned 169 speeds disk, supported by Data ONTAP 13 ways to mix disk, in aggregates 145 SSD storage pools commands for managing 143 creating Flash Pool aggregates using 173 See also storage pools SSDs adding to storage pools 141 capability differences with HDDs 26 changing size of RAID groups for 115 considerations for adding to existing storage pool versus new one 142 considerations for when to use storage pools 138 determining impact to cache size of adding to storage pool 143 how Data ONTAP manages wear life 26 how Data ONTAP reports disk types 10 how used in Flash Pool aggregates 130 introduction to using 25 shared See storage pools sizing considerations for RAID groups 114 storage pools, creating 140 storage pools, determining when used by a Flash Pool aggregate 193 SSL certificates installing on storage systems 87 installing replacement 99 preventing expiration issues 98 removing old 99 requirements 87 SSL connections introduction to using for secure key management communication 86 stacks configuring automatic ownership assignment for 61 standard layouts shared HDD 32 state of disks setting to end-of-life 103 stopping disk sanitization 51 storage adding to SSD storage pools 141 how root-data partitioning affects management of 31 how to control disk selection from heterogeneous 145 what happens when adding to an aggregate 170 storage aggregate commands for displaying space information 53 for managing aggregates 193 for managing disks 51 storage aggregate relocation start command key parameters of 187 storage arrays rules about mixing in aggregates 147 storage connection types supported 12 storage disk commands for managing disks 51 Storage Encryption benefits 83 destroying data using 102 displaying disk information 95 explained 81 how it works 82 information to collect before configuring 85 installing replacement SSL certificates 99 installing SSL certificates for 87 introduction to securing data at rest with 81 introduction to using SSL for secure key management communication 86 limitations 84 managing 90 overview 81 performing emergency data shredding 104 preventing SSL certificate expiration issues 98 purpose of external key management server 81 removing old SSL certificates 99 replacing self-encrypting disks 44 running the setup wizard 88 sanitizing disks using 102 setting disk state to end-of-life 103 setting up 85 SSL certificates requirements 87 storage limits aggregate 195 FlexClone file and LUN 195 RAID group 195 volume 195 storage performance 210 | Physical Storage Management Guide
introduction to using SSDs to increase 25 performance introduction to using SSDs to increase storage 25 storage pools adding SSDs to 141 advantages of SSD 138 commands for managing 143 considerations for adding SSDs to new versus existing 142 considerations for when to use SSD 138 creating 140 creating Flash Pool aggregates using SSD 173 determining impact to cache size of adding SSDs to 143 determining when used by a Flash Pool aggregate 193 disadvantages of SSD 138 how they increase cache allocation for Flash Pool aggregates 137 how you use 139 requirements and best practices for 139 storage shelves commands for displaying information about 52 requirements for using multi-disk carrier 27 storage systems considerations for removing disks from 45 suggestions how to send feedback about documentation 199 SVMs assigning aggregates to 190 effect on aggregate selection 144 SVMs with Infinite Volume aggregate requirements 152 assigning aggregates 154 delegation 154 SyncMirror protection with RAID and 110 system capacity methods of calculating 13 system performance impact of media scrubbing on 24 T terminology family 147 third-party storage verifying back-end configuration 72 timeouts RAID option, considerations for changing 121 tips for creating and backing up aggregates, for sensitive data 21 topologies supported storage connection type 12 twitter how to receive automatic notification of documentation changes 199 U unmirrored aggregates explained 127 unprotected mode returning SEDs to 100 unreachable key management servers precautions taken in the event of during boot process 94 used space how to determine and control in aggregates, by volume 150 how to determine in aggregate 148 V V-Series functionality name change to FlexArray Virtualization 64 V-Series systems See LUNs (array) verifying back-end configuration 72 key management server links 91 vetoes of an aggregate relocation 188 overriding 188 VMDK drives how Data ONTAP reports disk types 10 volume command for displaying space information 53 volume show-footprint command understanding output 150 volumes creating aggregates for FlexVol, using physical drives 159 creating aggregates for Infinite, using physical drives 159 determining which ones reside on an aggregate 192 determining write-caching eligibility for Flash Pool aggregates 178 Index | 211
how to determine space usage of, in aggregates 150 sharing of aggregates 152 See also Infinite Volumes Vservers See SVMs W WAFL external cache compared with Flash Pool aggregates 133 wear life how Data ONTAP manages SSD wear 26 wildcard characters how you use with disk ownership commands 62 wizards running the Storage Encryption setup 88 write caching determining Flash Pool aggregate eligibility 178 determining FlexVol volume eligibility 178 Z zoned checksum type changing for array LUNs 76 212 | Physical Storage Management Guide

More Related Content

PDF
ONTAP: Storage Management Guide
Kumaran Balachandran
 
PDF
DiskGenius User Guide
Eassos Ltd.
 
PDF
WHITE PAPER▶ Software Defined Storage at the Speed of Flash
Symantec
 
PDF
Platform Migration Guide
white paper
 
PDF
JM White
kunalboss
 
PDF
Virtual backup strategies_using_storage_snapshots_for_backups[1]
gerdev
 
PDF
My booklive manual
bb129207
 
PDF
8000 guide
suryakranthi27
 
ONTAP: Storage Management Guide
Kumaran Balachandran
 
DiskGenius User Guide
Eassos Ltd.
 
WHITE PAPER▶ Software Defined Storage at the Speed of Flash
Symantec
 
Platform Migration Guide
white paper
 
JM White
kunalboss
 
Virtual backup strategies_using_storage_snapshots_for_backups[1]
gerdev
 
My booklive manual
bb129207
 
8000 guide
suryakranthi27
 

What's hot (20)

PDF
Jasper server ce-install-guide
woid
 
PDF
ONTAP : System Administration Guide NetApp, Inc.
Kumaran Balachandran
 
PDF
Construction ofanoracle10glinuxserver 0.5
sopan sonar
 
PDF
Book.opensuse.startup en
Naldo Santos
 
PDF
Guide citrix presentation server™ - client for java administrator’s
xKinAnx
 
PDF
Rhel Tuningand Optimizationfor Oracle V11
Yusuf Hadiwinata Sutandar
 
PDF
Physical exercise web application documentation
TuononenP
 
PDF
820 6359-13
ramuktg
 
PDF
D space manual 1.5.2
tvcumet
 
PDF
Kenwood thd72e-manual-e
Diego Sirtori
 
PDF
Virtual box usermanual
fmartins_esas
 
PDF
Book VMWARE VMware ESXServer Advanced Technical Design Guide
aktivfinger
 
PDF
Data sharing planning and administration
Marino Savoldi
 
PDF
Video Jet 10 Manual
guest0ea020
 
PDF
CONTINUOUS SYSTEMS, NONSTOP OPERATIONS WITH JUNOS
Johnson Liu
 
PDF
Whats-New-VMware-vCloud-Director-15-Technical-Whitepaper
Djbilly Mixe Pour Toi
 
PDF
Vmware nsx-network-virtualization-platform-white-paper
CloudSyntrix
 
PDF
VMware-NSX-Network-Virtualization-Platform-WP
Stephen Fenton
 
PDF
Plesk 8.1 for Windows
webhostingguy
 
PDF
41450 04 rev-b_6gbs_megaraid_sas_raid_controllers_ug
sopan sonar
 
Jasper server ce-install-guide
woid
 
ONTAP : System Administration Guide NetApp, Inc.
Kumaran Balachandran
 
Construction ofanoracle10glinuxserver 0.5
sopan sonar
 
Book.opensuse.startup en
Naldo Santos
 
Guide citrix presentation server™ - client for java administrator’s
xKinAnx
 
Rhel Tuningand Optimizationfor Oracle V11
Yusuf Hadiwinata Sutandar
 
Physical exercise web application documentation
TuononenP
 
820 6359-13
ramuktg
 
D space manual 1.5.2
tvcumet
 
Kenwood thd72e-manual-e
Diego Sirtori
 
Virtual box usermanual
fmartins_esas
 
Book VMWARE VMware ESXServer Advanced Technical Design Guide
aktivfinger
 
Data sharing planning and administration
Marino Savoldi
 
Video Jet 10 Manual
guest0ea020
 
CONTINUOUS SYSTEMS, NONSTOP OPERATIONS WITH JUNOS
Johnson Liu
 
Whats-New-VMware-vCloud-Director-15-Technical-Whitepaper
Djbilly Mixe Pour Toi
 
Vmware nsx-network-virtualization-platform-white-paper
CloudSyntrix
 
VMware-NSX-Network-Virtualization-Platform-WP
Stephen Fenton
 
Plesk 8.1 for Windows
webhostingguy
 
41450 04 rev-b_6gbs_megaraid_sas_raid_controllers_ug
sopan sonar
 
Ad

Viewers also liked (14)

PDF
Data ONTAP 8.3: What's New?
Fast Lane Consulting and Education, Inc.
 
PDF
Basic concepts for_clustered_data_ontap_8.3_v1.1-lab_guide
Vikas Sharma
 
DOC
netapp c-mode terms
Ashwin Pawar
 
PDF
Taller de quimica organica segundo nivel 2017
sandro javier espin espin
 
PPTX
คำสั่งควบคุมโปรแกรม
BoOm mm
 
PDF
What's New in SecureAuth IdP in 2017
SecureAuth
 
PDF
Prezentare marta sms bucuresti 2017
RevistaBiz
 
PPTX
Finance in Dorset 2017 - Start Up Session
PKF Francis Clark
 
PPTX
Unit 3 - Egyptian art and architecture
JaimeAlonsoEdu
 
PDF
Lumière sur... le métier de conseiller communication ministériel
Damien ARNAUD
 
PDF
Lab thesis
Imran Nur Manik
 
PDF
Química orgánica SHT
sandro javier espin espin
 
PPTX
Question 7
Chris Wotton
 
PPTX
Mapa mental cooperativas
Ronny Ocanto
 
Data ONTAP 8.3: What's New?
Fast Lane Consulting and Education, Inc.
 
Basic concepts for_clustered_data_ontap_8.3_v1.1-lab_guide
Vikas Sharma
 
netapp c-mode terms
Ashwin Pawar
 
Taller de quimica organica segundo nivel 2017
sandro javier espin espin
 
คำสั่งควบคุมโปรแกรม
BoOm mm
 
What's New in SecureAuth IdP in 2017
SecureAuth
 
Prezentare marta sms bucuresti 2017
RevistaBiz
 
Finance in Dorset 2017 - Start Up Session
PKF Francis Clark
 
Unit 3 - Egyptian art and architecture
JaimeAlonsoEdu
 
Lumière sur... le métier de conseiller communication ministériel
Damien ARNAUD
 
Lab thesis
Imran Nur Manik
 
Química orgánica SHT
sandro javier espin espin
 
Question 7
Chris Wotton
 
Mapa mental cooperativas
Ronny Ocanto
 
Ad

Similar to Clustered data ontap_8.3_physical_storage (20)

PPTX
Netapp online training-30Hours Classes, 30Hours Lab,Assignments, Project
Vidhyalive
 
PDF
Net app
Acutelearn Technologies
 
PDF
Net app training at hyderabad
Acutelearn Technologies
 
PDF
Net app training in hyderabad
acute23
 
DOCX
How to assign unowned disk in the netapp cluster 8.3
Saroj Sahu
 
DOCX
NetApp system installation workbook Spokane
Accenture
 
PPTX
Geek Sync | Storage for the DBA
IDERA Software
 
PDF
10 Good Reasons to Upgrade to ONTAP 9.3
NetApp
 
PPT
Ds8000 Practical Performance Analysis P04 20060718
brettallison
 
PDF
Cr 2-2540 customer case study - How amgen utilizes Netapp clustered data Onta...
Harish Mundre
 
PDF
NetApp Clustered Data ONTAP Operating System and OnCommand Insight - Foundation
NetApp
 
PDF
NetAppOverview
madhu291670
 
PDF
NetApp IT Efficiencies Gained with Flash, NetApp ONTAP, OnCommand Insight, Al...
NetApp
 
PPTX
Accelerated NCDA Boot Camp Data ONTAP 7-Mode TrainingNetappp
royacuff .
 
PPTX
Macroview Netapp Overview
Alex Tsui
 
PDF
NetApp Training in Hyderabad
Acutelearn Technologies
 
PDF
Slides: Maintain 24/7 Availability for Your Enterprise Applications Environment
NetApp
 
PDF
Data ontap 8.x 7 mode cook book v1 1
Accenture
 
PDF
NetApp Administration and Best Practice, Brendon Higgins, Proact UK
subtitle
 
PPTX
NetApp C-mode for 7 mode engineers
subtitle
 
Netapp online training-30Hours Classes, 30Hours Lab,Assignments, Project
Vidhyalive
 
Net app
Acutelearn Technologies
 
Net app training at hyderabad
Acutelearn Technologies
 
Net app training in hyderabad
acute23
 
How to assign unowned disk in the netapp cluster 8.3
Saroj Sahu
 
NetApp system installation workbook Spokane
Accenture
 
Geek Sync | Storage for the DBA
IDERA Software
 
10 Good Reasons to Upgrade to ONTAP 9.3
NetApp
 
Ds8000 Practical Performance Analysis P04 20060718
brettallison
 
Cr 2-2540 customer case study - How amgen utilizes Netapp clustered data Onta...
Harish Mundre
 
NetApp Clustered Data ONTAP Operating System and OnCommand Insight - Foundation
NetApp
 
NetAppOverview
madhu291670
 
NetApp IT Efficiencies Gained with Flash, NetApp ONTAP, OnCommand Insight, Al...
NetApp
 
Accelerated NCDA Boot Camp Data ONTAP 7-Mode TrainingNetappp
royacuff .
 
Macroview Netapp Overview
Alex Tsui
 
NetApp Training in Hyderabad
Acutelearn Technologies
 
Slides: Maintain 24/7 Availability for Your Enterprise Applications Environment
NetApp
 
Data ontap 8.x 7 mode cook book v1 1
Accenture
 
NetApp Administration and Best Practice, Brendon Higgins, Proact UK
subtitle
 
NetApp C-mode for 7 mode engineers
subtitle
 

Recently uploaded (20)

PPTX
A Presentation on Artificial Intelligence
memembruh336
 
PDF
NewMind AI Weekly Chronicles - August'25 Week I
NewMind AI
 
PDF
Unlocking AI with Model Context Protocol (MCP)
Brian McKeiver
 
PDF
Peak of Data & AI Encore- AI for Metadata and Smarter Workflows
Safe Software
 
PPTX
VMware vSphere Foundation How to Sell Presentation-Ver1.4-2-14-2024.pptx
blackmambaettijean
 
PDF
Diabetes mellitus diagnosis method based random forest with bat algorithm
IAESIJAI
 
PDF
NewMind AI Monthly Chronicles - July 2025
NewMind AI
 
PDF
TokAI - TikTok AI Agent : The First AI Application That Analyzes 10,000+ Vira...
SOFTTECHHUB
 
PDF
CIFDAQ's Market Insight: SEC Turns Pro Crypto
CIFDAQ
 
PDF
Empathic Computing: Creating Shared Understanding
Mark Billinghurst
 
PDF
KodekX | Application Modernization Development
KodekX
 
PDF
The Rise and Fall of 3GPP – Time for a Sabbatical?
3G4G
 
PDF
Shreyas Phanse Resume: Experienced Backend Engineer | Java • Spring Boot • Ka...
SHREYAS PHANSE
 
PPTX
KOM of Painting work and Equipment Insulation REV00 update 25-dec.pptx
muhammadmuneebnaeem7
 
PDF
Per capita expenditure prediction using model stacking based on satellite ima...
IAESIJAI
 
PDF
7 ChatGPT Prompts to Help You Define Your Ideal Customer Profile.pdf
SOFTTECHHUB
 
PDF
Blue Purple Modern Animated Computer Science Presentation.pdf.pdf
Akar16
 
PDF
Reach Out and Touch Someone: Haptics and Empathic Computing
Mark Billinghurst
 
PDF
Agricultural_Statistics_at_a_Glance_2022_0.pdf
SandeepSingh286037
 
PPTX
Effective Security Operations Center (SOC) A Modern, Strategic, and Threat-In...
ReZa AdineH
 
A Presentation on Artificial Intelligence
memembruh336
 
NewMind AI Weekly Chronicles - August'25 Week I
NewMind AI
 
Unlocking AI with Model Context Protocol (MCP)
Brian McKeiver
 
Peak of Data & AI Encore- AI for Metadata and Smarter Workflows
Safe Software
 
VMware vSphere Foundation How to Sell Presentation-Ver1.4-2-14-2024.pptx
blackmambaettijean
 
Diabetes mellitus diagnosis method based random forest with bat algorithm
IAESIJAI
 
NewMind AI Monthly Chronicles - July 2025
NewMind AI
 
TokAI - TikTok AI Agent : The First AI Application That Analyzes 10,000+ Vira...
SOFTTECHHUB
 
CIFDAQ's Market Insight: SEC Turns Pro Crypto
CIFDAQ
 
Empathic Computing: Creating Shared Understanding
Mark Billinghurst
 
KodekX | Application Modernization Development
KodekX
 
The Rise and Fall of 3GPP – Time for a Sabbatical?
3G4G
 
Shreyas Phanse Resume: Experienced Backend Engineer | Java • Spring Boot • Ka...
SHREYAS PHANSE
 
KOM of Painting work and Equipment Insulation REV00 update 25-dec.pptx
muhammadmuneebnaeem7
 
Per capita expenditure prediction using model stacking based on satellite ima...
IAESIJAI
 
7 ChatGPT Prompts to Help You Define Your Ideal Customer Profile.pdf
SOFTTECHHUB
 
Blue Purple Modern Animated Computer Science Presentation.pdf.pdf
Akar16
 
Reach Out and Touch Someone: Haptics and Empathic Computing
Mark Billinghurst
 
Agricultural_Statistics_at_a_Glance_2022_0.pdf
SandeepSingh286037
 
Effective Security Operations Center (SOC) A Modern, Strategic, and Threat-In...
ReZa AdineH
 

Clustered data ontap_8.3_physical_storage

  • 1. Clustered Data ONTAP® 8.3 Physical Storage Management Guide NetApp, Inc. 495 East Java Drive Sunnyvale, CA 94089 U.S. Telephone: +1 (408) 822-6000 Fax: +1 (408) 822-4501 Support telephone: +1 (888) 463-8277 Web: www.netapp.com Feedback: doccomments@netapp.com Part number: 215-09163_B0 March 2015
  • 3. Contents Managing disks using Data ONTAP ......................................................... 10 How Data ONTAP reports disk types ....................................................................... 10 Storage connection types and topologies supported by Data ONTAP ...................... 12 How disks can be combined for the SAS storage connection type ............... 12 How disks can be combined for the FC-AL storage connection type ........... 12 Methods of calculating aggregate and system capacity ............................................ 13 Disk speeds supported by Data ONTAP ................................................................... 13 How drive checksum types affect aggregate and spare management ....................... 14 Drive name formats ................................................................................................... 14 Pre-cluster drive name formats ................................................................................. 15 Loop IDs for FC-AL connected disks ........................................................... 18 Understanding RAID drive types .............................................................................. 18 How disk sanitization works ..................................................................................... 19 Disk sanitization process ............................................................................... 19 When disk sanitization cannot be performed ................................................ 20 What happens if disk sanitization is interrupted ........................................... 20 Tips for creating and backing up aggregates containing data to be sanitized ................................................................................................... 21 How Data ONTAP monitors disk performance and health ....................................... 21 What happens when Data ONTAP takes disks offline ................................. 21 How Data ONTAP reduces disk failures using Rapid RAID Recovery ....... 21 How the maintenance center helps prevent drive errors ............................... 22 When Data ONTAP can put a disk into the maintenance center .................. 23 How Data ONTAP uses continuous media scrubbing to prevent media errors ........................................................................................................ 24 How you can use ACP to increase storage availability for SAS-connected disk shelves ................................................................................................................. 24 How you use SSDs to increase storage performance ................................................ 25 How the All-Flash Optimized personality affects node behavior ................. 25 How Data ONTAP manages SSD wear life .................................................. 26 Capability differences between SSDs and HDDs ......................................... 26 Guidelines and requirements for using multi-disk carrier storage shelves ............... 27 Table of Contents | 3
  • 4. How Data ONTAP avoids RAID impact when a multi-disk carrier must be removed .............................................................................................. 27 How to determine when it is safe to remove a multi-disk carrier ................. 28 Spare requirements for multi-disk carrier disks ............................................ 28 Shelf configuration requirements for multi-disk carrier storage shelves ...... 29 Aggregate requirements for disks in multi-disk carrier storage shelves ....... 29 Considerations for using disks from a multi-disk carrier storage shelf in an aggregate ............................................................................................. 29 Understanding root-data partitioning ........................................................................ 30 How root-data partitioning works ................................................................. 30 How root-data partitioning affects storage management .............................. 31 Which drives are partitioned and used for the root aggregate ....................... 32 Standard root-data partitioning layouts ......................................................... 32 Requirements for using root-data partitioning .............................................. 34 Initializing a node to configure root-data partitioning .................................. 34 Setting up an active-passive configuration on nodes using root-data partitioning .............................................................................................. 39 Adding disks to a node .............................................................................................. 41 When you need to update the Disk Qualification Package ........................... 42 Replacing disks that are currently being used in an aggregate ................................. 43 Replacing a self-encrypting disk ............................................................................... 44 Converting a data disk to a hot spare ........................................................................ 44 Removing disks from a node ..................................................................................... 45 Removing a failed disk .................................................................................. 45 Removing a hot spare disk ............................................................................ 46 Removing a data disk .................................................................................... 47 Using disk sanitization to remove data from disks ................................................... 48 Stopping disk sanitization ......................................................................................... 51 Commands for managing disks ................................................................................. 51 Commands for displaying information about storage shelves .................................. 52 Commands for displaying space usage information ................................................. 53 Managing ownership for disks .................................................................. 54 Types of disk ownership ........................................................................................... 54 Reasons to assign ownership of disks and array LUNs ............................................ 54 How disks and array LUNs become available for use .............................................. 55 How automatic ownership assignment works for disks ............................................ 56 4 | Physical Storage Management Guide
  • 5. Which disk autoassignment policy to use ..................................................... 56 When automatic ownership assignment is invoked ...................................... 57 How disk ownership works for platforms based on Data ONTAP-v technology ..... 57 Guidelines for assigning ownership for disks ........................................................... 58 Assigning ownership for disks .................................................................................. 58 Assigning ownership for disks partitioned for root-data partitioning ....................... 59 Removing ownership from a disk ............................................................................. 60 Configuring automatic ownership assignment of disks ............................................ 61 How you use the wildcard character with the disk ownership commands ................ 62 Managing array LUNs using Data ONTAP ............................................. 64 Data ONTAP systems that can use array LUNs on storage arrays ........................... 64 Installing the license for using array LUNs .............................................................. 65 How ownership for disks and array LUNs works ..................................................... 66 Reasons to assign ownership of disks and array LUNs ................................ 66 How disks and array LUNs become available for use .................................. 66 What it means for Data ONTAP to own an array LUN ................................ 68 Why you might assign array LUN ownership after installation .................... 68 Examples showing when Data ONTAP can use array LUNs ....................... 69 Assigning ownership of array LUNs ......................................................................... 71 Verifying the back-end configuration ....................................................................... 72 Modifying assignment of spare array LUNs ............................................................. 73 Array LUN name format ........................................................................................... 74 Pre-cluster array LUN name format .......................................................................... 74 Checking the checksum type of spare array LUNs ................................................... 76 Changing the checksum type of an array LUN ......................................................... 76 Prerequisites to reconfiguring an array LUN on the storage array ........................... 77 Changing array LUN size or composition ................................................................. 78 Removing one array LUN from use by Data ONTAP .............................................. 79 Preparing array LUNs before removing a Data ONTAP system from service ......... 79 Securing data at rest with Storage Encryption ........................................ 81 Introduction to Storage Encryption ........................................................................... 81 What Storage Encryption is ........................................................................... 81 How Storage Encryption works .................................................................... 82 Disk operations with SEDs ........................................................................... 82 Benefits of using Storage Encryption ............................................................ 83 Limitations of Storage Encryption ................................................................ 84 Table of Contents | 5
  • 6. Setting up Storage Encryption ................................................................................... 85 Information to collect before configuring Storage Encryption ..................... 85 Using SSL for secure key management communication .............................. 86 Running the Storage Encryption setup wizard .............................................. 88 Managing Storage Encryption ................................................................................... 90 Adding key management servers .................................................................. 90 Verifying key management server links ........................................................ 91 Displaying key management server information .......................................... 92 Removing key management servers .............................................................. 93 What happens when key management servers are not reachable during the boot process ....................................................................................... 94 Displaying Storage Encryption disk information .......................................... 95 Changing the authentication key ................................................................... 96 Retrieving authentication keys ...................................................................... 97 Deleting an authentication key ...................................................................... 98 SSL issues due to expired certificates ........................................................... 98 Returning SEDs to unprotected mode ......................................................... 100 Destroying data on disks using Storage Encryption ................................................ 102 Sanitizing disks using Storage Encryption before return to vendor ............ 102 Setting the state of disks using Storage Encryption to end-of-life .............. 103 Emergency shredding of data on disks using Storage Encryption .............. 104 What function the physical secure ID has for SEDs ................................... 105 SEDs that have PSID functionality ............................................................. 106 Resetting an SED to factory original settings ............................................. 106 How Data ONTAP uses RAID to protect your data and data availability ............................................................................................ 108 RAID protection levels for disks ............................................................................. 108 What RAID-DP protection is ...................................................................... 108 What RAID4 protection is ........................................................................... 109 RAID protection for array LUNs ............................................................................ 109 RAID protection for Data ONTAP-v storage ......................................................... 110 Protection provided by RAID and SyncMirror ....................................................... 110 Understanding RAID drive types ............................................................................ 113 How RAID groups work ......................................................................................... 113 How RAID groups are named ..................................................................... 114 Considerations for sizing RAID groups ...................................................... 114 6 | Physical Storage Management Guide
  • 7. Customizing the size of your RAID groups ................................................ 115 Considerations for Data ONTAP RAID groups for array LUNs ................ 116 How Data ONTAP works with hot spare disks ....................................................... 117 Minimum number of hot spares you should have ....................................... 117 What disks can be used as hot spares .......................................................... 118 What a matching spare is ............................................................................ 118 What an appropriate hot spare is ................................................................. 118 About degraded mode ................................................................................. 119 How low spare warnings can help you manage your spare drives .............. 120 How Data ONTAP handles a failed disk with a hot spare ...................................... 120 How Data ONTAP handles a failed disk that has no available hot spare ............... 121 Considerations for changing the timeout RAID option .......................................... 121 How RAID-level disk scrubs verify data integrity .................................................. 122 Changing the schedule for automatic RAID-level scrubs ........................... 122 How you run a manual RAID-level scrub ................................................... 123 Controlling the impact of RAID operations on system performance ...................... 123 Controlling the performance impact of RAID data reconstruction ............. 124 Controlling the performance impact of RAID-level scrubbing .................. 124 Controlling the performance impact of plex resynchronization .................. 125 Controlling the performance impact of mirror verification ........................ 126 What aggregates are ................................................................................. 127 How unmirrored aggregates work ........................................................................... 127 How mirrored aggregates work ............................................................................... 129 What a Flash Pool aggregate is ............................................................................... 130 How Flash Pool aggregates work ................................................................ 130 Requirements for using Flash Pool aggregates ........................................... 131 Considerations for RAID type and spare management for Flash Pool cache ...................................................................................................... 132 How Flash Pool aggregates and Flash Cache compare ............................... 133 How the available Flash Pool cache capacity is calculated ........................ 133 Using caching policies on volumes in Flash Pool aggregates ..................... 135 Modifying caching policies ......................................................................... 136 How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates .................................................................................................. 137 Considerations for when to use SSD storage pools ..................................... 138 How you use SSD storage pools ................................................................. 139 Table of Contents | 7
  • 8. Requirements and best practices for using SSD storage pools ................... 139 Creating an SSD storage pool ..................................................................... 140 Adding SSDs to an SSD storage pool ......................................................... 141 Considerations for adding SSDs to an existing storage pool versus creating a new one ................................................................................. 142 Determining the impact to cache size of adding SSDs to an SSD storage pool ........................................................................................................ 143 Commands for managing SSD storage pools .............................................. 143 How the SVM affects which aggregates can be associated with a FlexVol volume ............................................................................................................... 144 Understanding how Data ONTAP works with heterogeneous storage ................... 145 How you can use disks with mixed speeds in the same aggregate ............. 145 How to control disk selection from heterogeneous storage ........................ 145 Rules for mixing HDD types in aggregates ................................................ 146 Rules for mixing drive types in Flash Pool aggregates ............................... 147 Rules for mixing storage in array LUN aggregates ..................................... 147 How the checksum type is determined for array LUN aggregates ......................... 148 How to determine space usage in an aggregate ....................................................... 148 How you can determine and control a volume's space usage in the aggregate ....... 150 How Infinite Volumes use aggregates .................................................................... 152 Aggregate requirements for Infinite Volumes ............................................ 152 How FlexVol volumes and Infinite Volumes share aggregates .................. 152 How storage classes affect which aggregates can be associated with Infinite Volumes .................................................................................... 153 How aggregates and nodes are associated with Infinite Volumes .............. 154 How space is allocated inside a new Infinite Volume ................................ 155 Relocating ownership of aggregates used by Infinite Volumes .................. 156 Managing aggregates ............................................................................... 159 Creating an aggregate using unpartitioned drives ................................................... 159 Creating an aggregate using root-data partitioning ................................................. 161 Increasing the size of an aggregate that uses physical drives ................................. 163 Increasing the size of an aggregate that uses root-data partitioning ....................... 165 Correcting misaligned spare partitions .................................................................... 169 What happens when you add storage to an aggregate ............................................. 170 Creating a Flash Pool aggregate using physical SSDs ............................................ 171 Creating a Flash Pool aggregate using SSD storage pools ...................................... 173 8 | Physical Storage Management Guide
  • 9. Determining Flash Pool candidacy and optimal cache size .................................... 175 Determining and enabling volume write-caching eligibility for Flash Pool aggregates .......................................................................................................... 178 Changing the RAID type of RAID groups in a Flash Pool aggregate .................... 180 Determining drive and RAID group information for an aggregate ......................... 181 Relocating aggregate ownership within an HA pair ............................................... 182 How aggregate relocation works ................................................................. 183 How root-data partitioning affects aggregate relocation ............................. 184 Relocating aggregate ownership ................................................................. 184 Commands for aggregate relocation ........................................................... 187 Key parameters of the storage aggregate relocation start command ........... 187 Veto and destination checks during aggregate relocation ........................... 188 Assigning aggregates to SVMs ............................................................................... 190 Methods to create space in an aggregate ................................................................. 191 Determining which volumes reside on an aggregate .............................................. 192 Determining whether a Flash Pool aggregate is using an SSD storage pool .......... 193 Commands for managing aggregates ...................................................................... 193 Storage limits ............................................................................................ 195 Copyright information ............................................................................. 197 Trademark information ........................................................................... 198 How to send comments about documentation and receive update notification ............................................................................................ 199 Index ........................................................................................................... 200 Table of Contents | 9
  • 10. Managing disks using Data ONTAP Disks (sometimes also called drives) provide the basic unit of storage for storage systems running Data ONTAP that use native storage shelves. Understanding how Data ONTAP uses and classifies disks will help you manage your storage more effectively. How Data ONTAP reports disk types Data ONTAP associates a type with every disk. Data ONTAP reports some disk types differently than the industry standards; you should understand how Data ONTAP disk types map to industry standards to avoid confusion. When Data ONTAP documentation refers to a disk type, it is the type used by Data ONTAP unless otherwise specified. RAID disk types denote the role a specific disk plays for RAID. RAID disk types are not related to Data ONTAP disk types. For a specific configuration, the disk types supported depend on the storage system model, the shelf type, and the I/O modules installed in the system. The following tables show how Data ONTAP disk types map to industry standard disk types for the SAS and FC storage connection types, storage arrays, and for virtual storage (Data ONTAP-v). SAS-connected storage Data ONTAP disk type Disk class Industry standard disk type Description BSAS Capacity SATA Bridged SAS-SATA disks with added hardware to enable them to be plugged into a SAS-connected storage shelf. FSAS Capacity NL-SAS Near Line SAS MSATA Capacity SATA SATA disk in multi-disk carrier storage shelf SAS Performance SAS Serial-Attached SCSI SSD Ultra-performance SSD Solid-state drives 10 | Physical Storage Management Guide
  • 11. FC-connected storage Data ONTAP disk type Disk class Industry standard disk type Description ATA Capacity SATA FCAL Performance FC Storage arrays Data ONTAP disk type Disk class Industry standard disk type Description LUN N/A LUN A logical storage device backed by storage arrays and used by Data ONTAP as a disk. These LUNs are referred to as array LUNs to distinguish them from the LUNs that Data ONTAP serves to clients. Virtual storage (Data ONTAP-v) Data ONTAP disk type Disk class Industry standard disk type Description VMDISK N/A VMDK Virtual disks that are formatted and managed by VMware ESX. Related concepts Rules for mixing HDD types in aggregates on page 146 Storage connection types and topologies supported by Data ONTAP on page 12 Related references Understanding RAID drive types on page 18 Managing disks using Data ONTAP | 11
  • 12. Related information NetApp Technical Report 3437: Storage Subsystem Resiliency Guide NetApp Hardware Universe Storage connection types and topologies supported by Data ONTAP Data ONTAP supports two storage connection types: Serial-Attached SCSI (SAS) and Fibre Channel (FC). The FC connection type supports three topologies: arbitrated loop, switched, and point-to- point. • SAS, BSAS, FSAS, SSD, and MSATA disks use the SAS connection type. SAS-connected storage shelves are connected to the controller on a daisy chain called a stack. • FC and ATA disks use the FC connection type with an arbitrated-loop topology (FC-AL). FC-connected storage shelves are connected to the controller on a loop. • Array LUNs use the FC connection type, with either the point-to-point or switched topology. You cannot combine different connection types in the same loop or stack. However, for MetroCluster configurations, the FC and SAS connection types can be combined in a bridged connection, with FC on the controller side and SAS on the shelf side. The bridged connection can be used in either a direct-attached or switched topology. For more information, see Configuring a MetroCluster system with SAS disk shelves and FibreBridge 6500N bridges in 7-Mode. How disks can be combined for the SAS storage connection type You can combine SAS-connected storage shelves containing performance disks and SAS-connected storage shelves containing capacity disks within the same stack, although this configuration is not recommended. Each SAS-connected storage shelf can contain only one class of disk (capacity, performance, or SSDs). The only exception to this rule is if the shelf is being used for a Flash Pool aggregate. In that case, for some SSD sizes and shelf models, you can combine SSDs and HDDs in the same shelf. For more information, see the Hardware Universe. How disks can be combined for the FC-AL storage connection type You cannot combine storage shelves containing FC disks and storage shelves containing ATA disks in the same loop. 12 | Physical Storage Management Guide
  • 13. Methods of calculating aggregate and system capacity You use the physical and usable capacity of the drives you employ in your storage systems to ensure that your storage architecture conforms to the overall system capacity limits and the size limits of your aggregates. To maintain compatibility across different brands of drives, Data ONTAP rounds down (right-sizes) the amount of space available for user data. In addition, the numerical base used to calculate capacity (base 2 or base 10) also impacts sizing information. For these reasons, it is important to use the correct size measurement, depending on the task you want to accomplish: • For calculating overall system capacity, you use the physical capacity of the drive, and count every drive that is owned by the storage system. • For calculating how many drives you can put into an aggregate before you exceed its maximum size, you use the right-sized, or usable, capacity of all data drives in that aggregate. Parity, dparity, and cache drives are not counted against the maximum aggregate size. To see the physical and usable capacity for a specific drive, see the Hardware Universe at hwu.netapp.com. Disk speeds supported by Data ONTAP For hard disk drives, which use rotating media, speed is measured in revolutions per minute (RPM). Faster drives provide more input/output operations per second (IOPS) and faster response time. It is best to use disks of the same speed in an aggregate. Data ONTAP supports the following rotational speeds for hard disk drives: • Performance disks (SAS-connected) ◦ 10K RPM ◦ 15K RPM • Capacity disks (SAS-connected) ◦ 7.2K RPM • Performance disks (FC-connected) ◦ 15K RPM • Capacity disks (FC-connected) ◦ 7.2K RPM Managing disks using Data ONTAP | 13
  • 14. Solid-state drives, or SSDs, are flash media-based devices and therefore the concept of rotational speed does not apply to them. For more information about which disks are supported with specific hardware configurations, see the Hardware Universe at hwu.netapp.com. How drive checksum types affect aggregate and spare management There are two checksum types available for drives used by Data ONTAP: BCS (block) and AZCS (zoned). Understanding how the checksum types differ and how they impact storage management enables you to manage your storage more effectively. Both checksum types provide the same resiliency capabilities. BCS optimizes for data access speed, and reserves the smallest amount of capacity for the checksum for drives with 520-byte sectors. AZCS provides enhanced storage utilization and capacity for drives with 512-byte sectors. You cannot change the checksum type of a drive. To determine the checksum type of a specific drive model, see the Hardware Universe. Aggregates have a checksum type, which is determined by the checksum type of the drives or array LUNs that compose the aggregate. The following configuration rules apply to aggregates, drives, and checksums: • Checksum types cannot be combined within RAID groups. This means that you must consider checksum type when you provide hot spare drives. • When you add storage to an aggregate, if it has a different checksum type than the storage in the RAID group to which it would normally be added, Data ONTAP creates a new RAID group. • An aggregate can have RAID groups of both checksum types. These aggregates have a checksum type of mixed. • For mirrored aggregates, both plexes must have the same checksum type. • Drives of a different checksum type cannot be used to replace a failed drive. • You cannot change the checksum type of a drive. Drive name formats When a node is part of a functioning cluster, the drives it is connected to are accessible using a simple, consistent drive name format. The drive name is independent of what nodes the drive is physically connected to and from which node you are accessing the drive. The drive name format is a concatenation of four components: <stack_id>.<shelf_id>.<bay>.<position> 14 | Physical Storage Management Guide
  • 15. Note: During system boot, before the node has joined the cluster, or if certain key cluster components become unavailable, drive names revert to the classic format based on physical connectivity. • The stack ID is assigned by Data ONTAP. Stack IDs are unique across the cluster, and start with 1. • The shelf ID is set on the storage shelf when the shelf is added to the stack or loop. If there is a shelf ID conflict for SAS shelves, the shelf id is replaced with the shelf serial number in the drive name. • The bay is the position of the disk within its shelf. You can find the bay map in the administration guide for your storage shelf. • The position is used only for multi-disk carrier storage shelves. For carriers that house two disks, it can be 1 or 2. Related concepts Pre-cluster array LUN name format on page 74 Pre-cluster drive name formats on page 15 Pre-cluster drive name formats on page 15 Related references Commands for displaying information about storage shelves on page 52 Related information SAS Disk Shelves Installation and Service Guide for DS4243, DS2246, DS4486, and DS4246 DS14mk2 FC, and DS14mk4 FC Hardware Service Guide DiskShelf14mk2 AT Hardware Service Guide Pre-cluster drive name formats During steady-state operations, drive names are independent of their physical connection and unaffected by the node from which they are viewed. However, during system boot, before a node has joined a cluster, and if certain system components become unavailable, drive names revert to the format used before Data ONTAP 8.3, the pre-cluster format. The pre-cluster names of unowned drives (broken or unassigned drives) display the alphabetically lowest node name in the cluster that can see that drive. The following table shows the various formats for pre-cluster drive names, depending on how they are connected to the storage system. Managing disks using Data ONTAP | 15
  • 16. Note: For internal drives, the slot number is zero, and the internal port number depends on the system model. Drive connection Drive name Example SAS, direct-attached <node>:<slot><port>.<shelfID>.<bay> The pre-cluster name for the drive in shelf 2, bay 11, connected to onboard port 0a and owned by node1 is node1:0a. 2.11. The pre-cluster name for the drive in shelf 6, bay 3, connected to an HBA in slot 1, port c, and owned by node1 is node1:1c. 6.3. SAS, direct-attached in multi-disk carrier disk shelf <node>:<slot><port>.<shelfID>.<bay>L<carrierP osition> Carrier position is 1 or 2. SAS, direct-attached, for systems running Data ONTAP-v <slot><port>.<ID> The pre-cluster name for the third virtual disk connected to the first port is 0b.3. The pre-cluster name for the second virtual disk connected to the third port is 0d.2. The range of ports is b through e, and the range of disks is 0 through 15. 16 | Physical Storage Management Guide
  • 17. Drive connection Drive name Example SAS, bridge-attached (FibreBridge, used for MetroCluster configurations) <slot><port>.<loopID>L<LUN> The pre-cluster name for the drive with LUN 2 behind the bridge connected to port a in slot 3, loop ID 125, is 3a.125L2. SAS, bridge-attached through a switch (FibreBridge, used for MetroCluster configurations) <switch_name>:<switch_port>.<loopID>L<LUN > The pre-cluster name for the drive with LUN 5 behind the bridge connected to port 2 of switch brcd44, loop ID 126, is brcd44:2.126L5. FC, direct-attached <node>:<slot><port>.<loopID> The pre-cluster name for the drive with loop ID 19 (bay 3 of shelf 1) connected to onboard port 0a and owned by node1 is node1:0a. 19. The pre-cluster name for the drive with loop ID 34 connected to an HBA in slot 8, port c and owned by node1 is node1:8c. 34. FC, switch-attached <node>:<switch_name>.<switch_port>.<loopID> The pre-cluster name for the drive with loop ID 51 connected to port 3 of switch SW7 owned by node1 is node1:SW7.3.51. Managing disks using Data ONTAP | 17
  • 18. Each drive has a universal unique identifier (UUID) that differentiates it from all other drives in the cluster. Related concepts Drive name formats on page 14 Pre-cluster array LUN name format on page 74 Loop IDs for FC-AL connected disks For disks connected using Fibre Channel-Arbitrated Loop (FC-AL or FC), the loop ID is an integer between 16 and 126. The loop ID identifies the disk within its loop, and is included in the disk name, which identifies the disk uniquely for the entire system. The loop ID corresponds to the storage shelf number and the bay in which the disk is installed. The lowest loop ID is always in the far right bay of the first storage shelf. The next higher loop ID is in the next bay to the left, and so on. You can view the device map for your storage shelves by using the fcadmin device_map command, available through the nodeshell. For more information about the loop ID map for your storage shelf, see the hardware guide for the storage shelf. Understanding RAID drive types Data ONTAP classifies drives (or, for partitioned drives, partitions) as one of four types for RAID: data, hot spare, parity, or dParity. You manage disks differently depending on whether they are spare or being used in an aggregate. The RAID type is determined by how RAID is using a drive or partition; it is different from the Data ONTAP disk type. You cannot affect the RAID type for a drive. The RAID type is displayed in the Position column for many storage commands. For drives using root-data partitioning and SSDs in storage pools, a single drive might be used in multiple ways for RAID. For example, the root partition of a partitioned drive might be a spare partition, whereas the data partition might be being used for parity. For this reason, the RAID drive type for partitioned drives and SSDs in storage pools is displayed simply as shared. Data disk Holds data stored on behalf of clients within RAID groups (and any data generated about the state of the storage system as a result of a malfunction). Spare disk Does not hold usable data, but is available to be added to a RAID group in an aggregate. Any functioning disk that is not assigned to an aggregate but is assigned to a system functions as a hot spare disk. 18 | Physical Storage Management Guide
  • 19. Parity disk Stores row parity information that is used for data reconstruction when a single disk drive fails within the RAID group. dParity disk Stores diagonal parity information that is used for data reconstruction when two disk drives fail within the RAID group, if RAID-DP is enabled. Related concepts How Data ONTAP reports disk types on page 10 How disk sanitization works Disk sanitization is the process of physically obliterating data by overwriting disks or SSDs with specified byte patterns or random data so that recovery of the original data becomes impossible. You use the sanitization process to ensure that no one can recover the data on the disks. This functionality is available through the nodeshell. Related tasks Using disk sanitization to remove data from disks on page 48 Disk sanitization process Understanding the basics of the disk sanitization process helps you understand what to anticipate during the sanitization process and after it is complete. The disk sanitization process uses three successive default or user-specified byte overwrite patterns for up to seven cycles per operation. The random overwrite pattern is repeated for each cycle. Depending on the disk capacity, the patterns, and the number of cycles, the process can take several hours. Sanitization runs in the background. You can start, stop, and display the status of the sanitization process. The sanitization process contains two phases: 1. Formatting phase The operation performed for the formatting phase depends on the class of disk being sanitized, as shown in the following table: Disk class Formatting phase Capacity HDDs Skipped Performance HDDs SCSI format operation SSDs SCSI sanitize operation Managing disks using Data ONTAP | 19
  • 20. 2. Pattern overwrite phase The specified overwrite patterns are repeated for the specified number of cycles. When the sanitization process is complete, the specified disks are in a sanitized state. They are not returned to spare status automatically. You must return the sanitized disks to the spare pool before the newly sanitized disks are available to be added to another aggregate. When disk sanitization cannot be performed Disk sanitization is not supported for all disk types. In addition, there are times when disk sanitization cannot be performed. You should be aware of the following facts about the disk sanitization process: • It is not supported on all SSD part numbers. For information about which SSD part numbers support disk sanitization, see the Hardware Universe at hwu.netapp.com. • It is not supported in takeover mode for systems in an HA pair. • It cannot be performed on disks that were failed due to readability or writability problems. • It does not perform its formatting phase on ATA drives. • If you are using the random pattern, it cannot be performed on more than 100 disks at one time. • It is not supported on array LUNs. • If you sanitize both SES disks in the same ESH shelf at the same time, you see errors on the console about access to that shelf, and shelf warnings are not reported for the duration of the sanitization. However, data access to that shelf is not interrupted. • You can perform disk sanitization on disks using Storage Encryption. However, there are other methods to obliterate data on disks using Storage Encryption that are faster and do not require an operational storage system. What happens if disk sanitization is interrupted Disk sanitization is a long-running operation. If disk sanitization is interrupted by user intervention or an unexpected event such as a power outage, Data ONTAP takes action to return the disks that were being sanitized to a known state, but you must also take action before the sanitization process can finish. If the sanitization process is interrupted by power failure, system panic, or manual intervention, the sanitization process must be repeated from the beginning. The disk is not designated as sanitized. If the formatting phase of disk sanitization is interrupted, Data ONTAP must recover any disks that were corrupted by the interruption. After a system reboot and once every hour, Data ONTAP checks for any sanitization target disk that did not complete the formatting phase of its sanitization. If any 20 | Physical Storage Management Guide
  • 21. such disks are found, Data ONTAP recovers them. The recovery method depends on the type of the disk. After a disk is recovered, you can rerun the sanitization process on that disk; for HDDs, you can use the -s option to specify that the formatting phase is not repeated again. Tips for creating and backing up aggregates containing data to be sanitized If you are creating or backing up aggregates to contain data that might need to be sanitized, following some simple guidelines will reduce the time it takes to sanitize your data. • Make sure your aggregates containing sensitive data are not larger than they need to be. If they are larger than needed, sanitization requires more time, disk space, and bandwidth. • When you back up aggregates containing sensitive data, avoid backing them up to aggregates that also contain large amounts of nonsensitive data. This reduces the resources required to move nonsensitive data before sanitizing sensitive data. How Data ONTAP monitors disk performance and health Data ONTAP continually monitors disks to assess their performance and health. When Data ONTAP encounters certain errors or behaviors from a disk, it takes the disk offline temporarily or takes the disk out of service to run further tests. What happens when Data ONTAP takes disks offline Data ONTAP temporarily stops I/O activity to a disk and takes a disk offline when Data ONTAP is updating disk firmware in background mode or when disks become non-responsive. While the disk is offline, Data ONTAP performs a quick check on it to reduce the likelihood of forced disk failures. A disk can be taken offline only if its containing RAID group is in a normal state and the plex or aggregate is not offline. While the disk is offline, Data ONTAP reads from other disks within the RAID group while writes are logged. When the offline disk is ready to come back online, Data ONTAP resynchronizes the RAID group and brings the disk online. This process generally takes a few minutes and incurs a negligible performance impact. How Data ONTAP reduces disk failures using Rapid RAID Recovery When Data ONTAP determines that a disk has exceeded its error thresholds, Data ONTAP can perform Rapid RAID Recovery by removing the disk from its RAID group for testing and, if necessary, failing the disk. Spotting disk errors quickly helps prevent multiple disk failures and allows problem disks to be replaced. By performing the Rapid RAID Recovery process on a suspect disk, Data ONTAP avoids three problems that occur during sudden disk failure and the subsequent RAID reconstruction process: • Rebuild time Managing disks using Data ONTAP | 21
  • 22. • Performance degradation • Potential data loss due to additional disk failure during reconstruction During Rapid RAID Recovery, Data ONTAP performs the following tasks: 1. Places the suspect disk in pre-fail mode. 2. Selects a hot spare replacement disk. Note: If no appropriate hot spare is available, the suspect disk remains in pre-fail mode and data continues to be served. However, a suspect disk performs less efficiently. Impact on performance ranges from negligible to worse than degraded mode. For this reason, hot spares should always be available. 3. Copies the suspect disk’s contents to the spare disk on the storage system before an actual failure occurs. 4. After the copy is complete, attempts to put the suspect disk into the maintenance center, or else fails the disk. Note: Tasks 2 through 4 can occur only when the RAID group is in normal (not degraded) mode. If the suspect disk fails on its own before copying to a hot spare is complete, Data ONTAP starts the normal RAID reconstruction process. A message is sent to the log file when the Rapid RAID Recovery process is started and when it is complete. The messages are tagged raid.rg.diskcopy.start:notice and raid.rg.diskcopy.done:notice. Related concepts About degraded mode on page 119 When Data ONTAP can put a disk into the maintenance center on page 23 How Data ONTAP works with hot spare disks on page 117 How the maintenance center helps prevent drive errors Drives can sometimes display small problems that do not interfere with normal operation, but which could be a sign that the drive might fail sometime soon. The maintenance center provides a way to put these drives under increased scrutiny. When a suspect drive is in the maintenance center, it is subjected to a number of tests. If the drive passes all of the tests, Data ONTAP redesignates it as a spare; if it fails any tests, Data ONTAP fails the drive. By default, Data ONTAP puts a suspect drive into the maintenance center automatically only if there are two or more spares available for that drive. If that drive is housed in a shelf that supports automatic power-cycling, power to that drive might be turned off for a short time. If the drive returns 22 | Physical Storage Management Guide
  • 23. to a ready state after the power-cycle, the maintenance center tests the drive. Otherwise, the maintenance center fails the drive immediately. You can put a suspect drive into the maintenance center manually regardless of how many spares are available. You can also specify the number of times a drive is allowed to go to the maintenance center and whether it should go immediately or only after copying its contents to a spare drive. • The disk.maint_center.enable option controls whether the maintenance center is on or off. The default value is on. • The disk.maint_center.allowed_entries option controls the number of times a suspect drive is allowed to go to the maintenance center. The default value is 1, which means that if the drive is sent to the maintenance center more than once, it is automatically failed. • The disk.maint_center.spares_check option controls whether Data ONTAP requires that sufficient spares are available before it puts a drive into the maintenance center. • The disk maint start command, available through the nodeshell, enables you to put a suspect drive into the maintenance center manually. If the target drive is in use, the default is that it does not enter the maintenance center until its contents have been copied to a spare drive. You can put the drive into the maintenance center immediately by including the -i option. Related information NetApp Technical Report 3437: Storage Subsystem Resiliency Guide When Data ONTAP can put a disk into the maintenance center When Data ONTAP detects certain disk errors, it tries to put the disk into the maintenance center for testing. Certain requirements must be met for the disk to be put into the maintenance center. If a disk experiences more errors than are allowed for that disk type, Data ONTAP takes one of the following actions: • If the disk.maint_center.spares_check option is set to on (the default) and two or more spares are available (four for multi-disk carriers), Data ONTAP takes the disk out of service and assigns it to the maintenance center for data management operations and further testing. • If the disk.maint_center.spares_check option is set to on and fewer than two spares are available (four for multi-disk carriers), Data ONTAP does not assign the disk to the maintenance center. It fails the disk and designates the disk as a broken disk. • If the disk.maint_center.spares_check option is set to off, Data ONTAP assigns the disk to the maintenance center without checking the number of available spares. Managing disks using Data ONTAP | 23
  • 24. Note: The disk.maint_center.spares_check option has no effect on putting disks into the maintenance center from the command-line interface. Data ONTAP does not put SSDs into the maintenance center. How Data ONTAP uses continuous media scrubbing to prevent media errors The purpose of the continuous media scrub is to detect and correct media errors to minimize the chance of storage system disruption due to a media error while a storage system is in degraded or reconstruction mode. By default, Data ONTAP runs continuous background media scrubbing for media errors on all storage system disks. If a media error is found, Data ONTAP uses RAID to reconstruct the data and repairs the error. Media scrubbing is a continuous background process. Therefore, you might observe disk LEDs blinking on an apparently idle storage system. You might also observe some CPU activity even when no user workload is present. How continuous media scrubbing impacts system performance Because continuous media scrubbing searches only for media errors, its impact on system performance is negligible. In addition, the media scrub attempts to exploit idle disk bandwidth and free CPU cycles to make faster progress. However, any client workload results in aggressive throttling of the media scrub resource. If needed, you can further decrease the CPU resources consumed by a continuous media scrub under a heavy client workload by increasing the maximum time allowed for a media scrub cycle to complete. You can do this by using the raid.media_scrub.rate option. Why continuous media scrubbing should not replace scheduled RAID-level disk scrubs Because the continuous media scrub process scrubs only media errors, you should continue to run the storage system’s scheduled complete RAID-level scrub operation. The RAID-level scrub finds and corrects parity and checksum errors as well as media errors. How you can use ACP to increase storage availability for SAS-connected disk shelves The Alternate Control Path (ACP) protocol enables Data ONTAP to manage and control SAS- connected storage shelves, which can increase storage availability. After you ensure that ACP is properly enabled and configured, you do not need to actively manage it. You can install SAS-connected storage shelves without configuring ACP. However, for maximum storage availability and stability, you should always have ACP enabled and configured, including 24 | Physical Storage Management Guide
  • 25. providing the required physical connectivity and configuration parameters to enable the ACP functionality. ACP is enabled by default. If you need to enable it or change its configuration, you can use the acpadmin configure and storage show acp commands, available through the nodeshell. You do not need to actively manage the SAS-connected storage shelf subsystem. Data ONTAP automatically monitors and manages the subsystem without operator intervention. How you use SSDs to increase storage performance Solid-state drives (SSDs) are flash media-based storage devices that provide better overall performance than hard disk drives (HDDs), which are mechanical devices using rotating media. You should understand how Data ONTAP manages SSDs and the capability differences between SSDs and HDDs. Depending on your storage system model, you can use SSDs in two ways: • You can create Flash Pool aggregates—aggregates composed mostly of HDDs, but with some SSDs that function as a high-performance cache for your working data set. • You can create aggregates composed entirely of SSDs, where the SSDs function as the persistent storage for all data in the aggregate. You manage Flash Pool aggregates and aggregates composed entirely of SSDs the same way you manage aggregates composed entirely of HDDs. However, there are some differences in the way you manage SSDs from the way you manage disks. In addition, some Data ONTAP capabilities are not available on SSDs and Flash Pool aggregates. Related concepts How Flash Pool aggregates work on page 130 How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates on page 137 How the All-Flash Optimized personality affects node behavior When you order an All-Flash FAS platform model, it has the All-Flash Optimized personality. Having the All-Flash Optimized personality enabled for a node introduces some extra requirements for that node. When the All-Flash Optimized personality is enabled, you cannot attach HDDs or array LUNs to the node; no drives other than SSDs are recognized as supported. This enables the node to provide performance that can be achieved only with a pure SSD solution. Both partners in an HA pair must have the same All-Flash Optimized personality (either enabled or disabled). You cannot move an aggregate that contains HDD RAID groups to a node with the All- Managing disks using Data ONTAP | 25
  • 26. Flash Optimized personality. Volume move is unaffected; you can move volumes between nodes with and nodes without the All-Flash Optimized personality. Root-data partitioning is enabled by default on any node with the All-Flash Optimized personality. You can determine whether a node has the All-Flash Optimized personality by using the system node show command and looking for All-Flash Optimized. How Data ONTAP manages SSD wear life Solid-state disks (SSDs) have a different end-of-life behavior than rotating media (hard disk drives, or HDDs). Data ONTAP monitors and manages SSDs to maximize storage performance and availability. In the absence of a mechanical failure, rotating media can serve data almost indefinitely. This is not true for SSDs, which can accept only a finite (though very large) number of write operations. SSDs provide a set of internal spare capacity, called spare blocks, that can be used to replace blocks that have reached their write operation limit. After all of the spare blocks have been used, the next block that reaches its limit causes the disk to fail. Because a drive failure is an undesirable occurrence, Data ONTAP replaces SSDs before they reach their limit. When a predetermined percentage of the spare blocks have been used (approximately 90%), Data ONTAP performs the following actions: 1. Sends an AutoSupport message. 2. If a spare SSD is available, starts a disk copy to that spare. 3. If no spare is available, starts a periodic check for a spare so that the disk copy can be started when a spare becomes available. 4. When the disk copy finishes, fails the disk. Note: You do not need to replace SSDs before they are failed by Data ONTAP. However, when you use SSDs in your storage system (as for all disk types), it is important to ensure that you have sufficient hot spares available at all times. Capability differences between SSDs and HDDs Usually, you manage SSDs the same as HDDs, including firmware updates, scrubs, and zeroing. However, some Data ONTAP capabilities do not make sense for SSDs, and SSDs are not supported on all hardware models. SSDs cannot be combined with HDDs within the same RAID group. When you replace an SSD in an aggregate, you must replace it with another SSD. Similarly, when you physically replace an SSD within a shelf, you must replace it with another SSD. The following capabilities of Data ONTAP are not available for SSDs: • Disk sanitization is not supported for all SSD part numbers. 26 | Physical Storage Management Guide
  • 27. • The maintenance center Guidelines and requirements for using multi-disk carrier storage shelves Data ONTAP automatically handles most of the extra steps required to manage disks in multi-disk carriers. However, there are some extra management and configuration requirements that you must understand before incorporating multi-disk carrier disk shelves in your storage architecture. When using storage from multi-disk carrier disk shelves such as the DS4486, you must familiarize yourself with the guidelines and requirements governing the following topics: • The process that Data ONTAP uses to avoid impacting any RAID groups when a multi-disk carrier needs to be removed • When it is safe to remove a multi-disk carrier after a disk failure • The minimum required number of spares for multi-disk carrier disks • Multi-disk carrier disk shelf configuration • Aggregate configuration requirements when using multi-disk carrier disk shelves • Guidelines and best practices for using disks from a multi-disk carrier disk shelf in an aggregate How Data ONTAP avoids RAID impact when a multi-disk carrier must be removed Data ONTAP takes extra steps to ensure that both disks in a carrier can be replaced without impacting any RAID group. Understanding this process helps you know what to expect when a disk from a multi-disk carrier storage shelf fails. A multi-disk carrier storage shelf, such as the DS4486, has double the storage density of other SAS- connected storage shelves. It accomplishes this by housing two disks per disk carrier. When two disks share the same disk carrier, they must be removed and inserted together. This means that when one of the disks in a carrier needs to be replaced, the other disk in the carrier must also be replaced, even if it was not experiencing any issues. Removing two data or parity disks from an aggregate at the same time is undesirable, because it could leave two RAID groups degraded, or one RAID group double-degraded. To avoid this situation, Data ONTAP initiates a storage evacuation operation for the carrier mate of the failed disk, as well as the usual reconstruction to replace the failed disk. The disk evacuation operation copies the contents of the carrier mate to a disk in a different carrier so that the data on that disk remains available when you remove the carrier. During the evacuation operation, the status for the disk being evacuated is shown as evacuating. In addition, Data ONTAP tries to create an optimal layout that avoids having two carrier mates in the same RAID group. Depending on how the other disks are laid out, achieving the optimal layout can Managing disks using Data ONTAP | 27
  • 28. require as many as three consecutive disk evacuation operations. Depending on the size of the disks and the storage system load, each storage evacuation operation could take several hours, so the entire swapping process could take an entire day or more. If insufficient spares are available to support the swapping operation, Data ONTAP issues a warning and waits to perform the swap until you provide enough spares. How to determine when it is safe to remove a multi-disk carrier Removing a multi-disk carrier before it is safe to do so can result in one or more RAID groups becoming degraded, or possibly even a storage disruption. Data ONTAP provides several indications of when it is safe to remove a multi-disk carrier. When a multi-disk carrier needs to be replaced, the following events must have occurred before you can remove the carrier safely: • An AutoSupport message must have been logged indicating that the carrier is ready to be removed. • An EMS message must have been logged indicating that the carrier is ready to be removed. • Both disks in the carrier must be displayed in the list of broken disks. You can see the list of broken disks by using the storage disk show -broken command. The disk that was evacuated to allow the carrier to be removed shows the outage reason of evacuated. • The amber LED on the carrier must be lit continuously. • The green LED on the carrier must show no activity. Attention: You cannot reuse the carrier mate of a failed disk. When you remove a multi-disk carrier that contains a failed disk, you must replace and return the entire carrier. Spare requirements for multi-disk carrier disks Maintaining the proper number of spares for disks in multi-disk carriers is critical for optimizing storage redundancy and minimizing the amount of time Data ONTAP must spend copying disks to achieve an optimal disk layout. You must maintain a minimum of two hot spares for multi-disk carrier disks at all times. To support the use of the Maintenance Center, and to avoid issues caused by multiple concurrent disk failures, you should maintain at least four hot spares for steady state operation, and replace failed disks promptly. If two disks fail at the same time with only two available hot spares, Data ONTAP might not be able to swap the contents of both the failed disk and its carrier mate to the spare disks. This scenario is called a stalemate. If this happens, you are notified through EMS messages and AutoSupport messages. When the replacement carriers become available, you must follow the instructions provided by the EMS messages or contact technical support to recover from the stalemate. 28 | Physical Storage Management Guide
  • 29. Shelf configuration requirements for multi-disk carrier storage shelves You can combine multi-disk carrier disk shelves with single-disk carrier disk shelves (standard disk shelves) on the same storage system and within in the same stack. Aggregate requirements for disks in multi-disk carrier storage shelves Aggregates composed of disks in multi-disk carrier disk shelves must conform to some configuration requirements. The following configuration requirements apply to aggregates composed of disks in multi-disk carrier disk shelves: • The RAID type must be RAID-DP. • All HDDs in the aggregate must be the same Data ONTAP disk type. The aggregate can be a Flash Pool aggregate. • If the aggregate is mirrored, both plexes must have the same Data ONTAP disk type (or types, in the case of a Flash Pool aggregate). Related concepts How Flash Pool aggregates work on page 130 Considerations for using disks from a multi-disk carrier storage shelf in an aggregate Observing the requirements and best practices for using disks from a multi-disk carrier disk shelf in an aggregate enables you to maximize storage redundancy and minimize the impact of disk failures. Disks in multi-disk carriers always have the Data ONTAP disk type of MSATA. MSATA disks cannot be mixed with HDDs from a single-carrier disk shelf in the same aggregate. The following disk layout requirements apply when you are creating or increasing the size of an aggregate composed of MSATA disks: • Data ONTAP prevents you from putting two disks in the same carrier into the same RAID group. • Do not put two disks in the same carrier into different pools, even if the shelf is supplying disks to both pools. • Do not assign disks in the same carrier to different nodes. • For the best layout, do not name specific disks; allow Data ONTAP to select the disks to be used or added. If the operation cannot result in an optimal layout due to the placement of the disks and available spares, Data ONTAP automatically swaps disk contents until an optimal layout is achieved. If there are not enough available spares to support the swaps, Data ONTAP issues a warning and Managing disks using Data ONTAP | 29
  • 30. waits to perform the disk swaps until you provide the necessary number of hot spares. If you name disks and an optimal layout cannot be achieved, you must explicitly force the operation; otherwise, the operation fails. Aggregate creation example To create an aggregate using MSATA disks, you can specify the disk type and size but leave the disk selection and layout to Data ONTAP by using a command like this: storage aggregate create -aggregate c1n1_aggr1 -node node1 - disktype MSATA -diskcount 14 Understanding root-data partitioning Some platform models use partioning to enable the root aggregate to use less space, leaving more space for the data aggregate and improving storage utilization. Root-data partitioning, also called shared drives, changes the way you view and administer your disks and aggregates. How root-data partitioning works For entry-level and All Flash FAS (AFF) platform models, aggregates can be composed of parts of a drive rather than the entire drive. Root-data partitioning is usually enabled and configured by the factory. It can also be established by initiating system initialization using option 4 from the boot menu. Note that system initialization erases all data on the disks of the node and resets the node configuration to the factory default settings. When a node has been configured to use root-data partitioning, partitioned disks have two partitions: The smaller partition is used to compose the root aggregate. The larger partition is used in data aggregates. The size of the partitions is set by Data ONTAP, and depends on the number of disks used to compose the root aggregate when the system is initialized. (The more disks used to compose the root aggregate, the smaller the root partition.) After system initialization, the partition sizes are fixed; adding partitions or disks to the root aggregate after system initialization increases the size of the root aggregate, but does not change the root partition size. 30 | Physical Storage Management Guide
  • 31. The partitions are used by RAID in the same manner as physical disks are; all of the same requirements apply. For example, if you add an unpartitioned drive to a RAID group consisting of partitioned drives, the unpartitioned drive is partitioned to match the partition size of the drives in the RAID group and the rest of the disk is unused. If a partitioned disk is moved to another node or used in another aggregate, the partitioning persists; you can use the disk only in RAID groups composed of partitioned disks. Related information Clustered Data ONTAP 8.3 System Administration Guide for Cluster Administrators How root-data partitioning affects storage management Generally, you administer a node that is using root-data partitioning the same way you administer a node that uses only physical disks. However, partitioned disks are displayed slightly differently, and you need to take their presence into account for some management tasks. Root and data partitions are used by the RAID subsystem the same way that physical disks are, and the same requirements and best practices apply. You follow the same best practices for spare management, and disk names are not impacted by root-data partitioning. When you are working with a node that is using root-data partitioning, the following tasks are slightly different: • Creating and adding storage to aggregates Spare partitions are displayed differently—as either root or data. In addition, you need to ensure that a root and data spare are provided on the same disk to support core file creation. Finally, you should understand whether you are mixing physical disks and partitions in the same RAID group, and the impact of doing so. • Displaying drive and RAID group information for an aggregate Partitioned disks display shared for the Position column, rather than their RAID disk type. • Assigning and removing ownership Automatic ownership assignment works for partitioned disks. If you need to manually assign or remove ownership for partitioned disks, you must do so for both partitions as well as the container disk. • Removing and replacing disks Removing a partitioned disk can potentially affect more than one aggregate. Related concepts Which drives are partitioned and used for the root aggregate on page 32 Removing disks from a node on page 45 Requirements for using root-data partitioning on page 34 Managing disks using Data ONTAP | 31
  • 32. Related tasks Creating an aggregate using root-data partitioning on page 161 Increasing the size of an aggregate that uses root-data partitioning on page 165 Determining drive and RAID group information for an aggregate on page 181 Assigning ownership for disks partitioned for root-data partitioning on page 59 Related references Commands for managing disks on page 51 Which drives are partitioned and used for the root aggregate The drives that are partitioned for use in the root aggregate depend on the system configuration. Knowing how many drives are used for the root aggregate helps you to determine how much of the drives' capacity is reserved for the root partition, and how much is available for use in a data aggregate. Root-data partitioning is supported for two types of platform: entry-level platforms, and All-Flash FAS (AFF) platforms. For entry-level platforms, only the internal drives are partitioned. For AFF platforms, all drives that are attached to the controller when the system is initialized are partitioned, up to a limit of 24 per node. Drives that are added after system configuration are not partitioned. Standard root-data partitioning layouts The root aggregate is configured by the factory; you should not change it. However, you can use the data partitions in a few different configurations, depending on your requirements. The following diagram shows one way to configure the partitions for an active-passive configuration with 12 partitioned disks. There are two root aggregates, one for each node, composed of the small partitions. Each root aggregate has a spare partition. There is just one RAID-DP data aggregate, with two parity disk partitions and one spare partition. 32 | Physical Storage Management Guide
  • 33. The following diagram shows one way to configure the partitions for an active-active configuration with 12 partitioned disks. In this case, there are two RAID-DP data aggregates, each with their own data partitions, parity partitions, and spares. Note that each disk is allocated to only one node. This is a best practice that prevents the loss of a single disk from affecting both nodes. The disks used for data, parity, and spare partitions might not be exactly as shown in these diagrams. For example, the parity partitions might not always align on the same disk. Managing disks using Data ONTAP | 33
  • 34. Requirements for using root-data partitioning In most cases, you can use drives that are partitioned for root-data partitioning exactly as you would use a physical, unshared drive. However, you cannot use root-data partitioning in certain configurations. The following storage devices cannot be partitioned: • Array LUNs • Virtual disks as created by Data ONTAP-v • HDD types that are not available as internal drives: ATA, FCAL, and MSATA • 100-GB SSDs You cannot use root-data partitioning with the following technologies: • MetroCluster • Data ONTAP-v • RAID4 Aggregates composed of partitioned drives must have a RAID type of RAID-DP. Initializing a node to configure root-data partitioning If you are initializing a node whose platform model supports root-data partitioning, you must complete some steps before performing the initialization to ensure that root-data partitioning is configured correctly. Before you begin • Attention: All data that you need to retain must have been migrated to another node. All data on the HA pair is erased during this procedure and cannot be retrieved. • Your node configuration must support root-data partitioning. If the node does not satisfy the requirements for root-data partitioning, root-data partitioning is not configured. • The node or HA pair you are initializing should not be joined with more nodes in a cluster, and should not be serving data. • Your nodes must be running Data ONTAP 8.3 or later. • Neither node is in takeover or giveback. 34 | Physical Storage Management Guide
  • 35. About this task This procedure can be used for both entry-level platform models and All-Flash FAS (AFF) platform models. For entry-level platform models, only internal disks are partitioned. For AFF platform models, up to 48 SSDs are partitioned, depending on the number of SSDs that are attached to the controller. This procedure is designed to be run on an HA pair, so the nodes are called Node A and Node B. However, you can still use this procedure if you have only one controller; ignore instructions that are specifically for Node B. This procedure can take many hours to complete, depending on the amount and type of storage attached to the HA pair. Steps 1. Record your node configuration, including network configuration, license values, and passwords, and if you want to restore the same storage architecture as you have now, record that also. All of the current node configuration information will be erased when the system is initialized. 2. If you are performing this procedure on a two-node cluster, disable the HA capability: cluster ha modify -configured false This can be done from either node. 3. Disable storage failover for both nodes: storage failover modify -enabled false -node nodeA,nodeB 4. On both nodes, boot into maintenance mode: a. Halt the system: system node halt -node node_name You can ignore any error messages about epsilon. For the second node, you can include the - skip-lif-migration-before-shutdown flag if prompted to do so. b. At the LOADER prompt, boot Data ONTAP: boot_ontap c. Monitor the boot process, and when prompted, press Ctrl-C to display the boot menu. Managing disks using Data ONTAP | 35
  • 36. Example ******************************* * * * Press Ctrl-C for Boot Menu. * * * ******************************* d. Select option 5, Maintenance mode boot. Example Please choose one of the following: (1) Normal Boot. (2) Boot without /etc/rc. (3) Change password. (4) Clean configuration and initialize all disks. (5) Maintenance mode boot. (6) Update flash from backup config. (7) Install new software first. (8) Reboot node. Selection (1-8)? 5 5. For both nodes, if there are external disks connected to the node, destroy all aggregates, including the root aggregate: a. Display all aggregates: aggr status b. For each aggregate, take the aggregate offline: aggr offline aggr_name c. For each aggregate, destroy the aggregate: aggr destroy aggr_name All disks are converted to spares. 6. Ensure that no drives in the HA pair are partitioned: a. Display all drives owned by both nodes: disk show If any of the drives show three entries, that drive has been partitioned. If your nodes have no partitioned disks, you can skip to step 7. The following example shows partitioned drive 0a.10.11: 36 | Physical Storage Management Guide
  • 37. Example DISK OWNER POOL SERIAL NUMBER HOME ------------ ------------- ----- ------------- ------------- 0a.10.11 sys1(536880559) Pool0 N11YE08L sys1(536880559) 0a.10.11P1 sys1(536880559) Pool0 N11YE08LNP001 sys1(536880559) 0a.10.11P2 sys1(536880559) Pool0 N11YE08LNP002 sys1(536880559) b. Note any partitions that do not have the same owner as their container disk. Example In the following example, the container disk (0a.10.14) is owned by sys1, but partition one (0a.10.14P1) is owned by sys2. DISK OWNER POOL SERIAL NUMBER HOME ------------ ------------- ----- ------------- ------------- 0a.10.14 sys1(536880559) Pool0 N11YE08L sys1(536880559) 0a.10.14P1 sys2(536880408) Pool0 N11YE08LNP001 sys2(536880408) 0a.10.14P2 sys1(536880559) Pool0 N11YE08LNP002 sys1(536880559) c. Update the ownership of all partitions owned by a different node than their container disk: disk assign disk_partition -f -o container_disk_owner Example You would enter a command like the following example for each disk with partitions owned by a different node than their container disk: disk assign 0a.10.14P1 -f -o sys1 d. For each drive with partitions, on the node that owns the container disk, remove the partitions: disk unpartition disk_name disk_name is the name of the disk, without any partition information, such as “0a.10.3”. Example You would enter a command like the following example for each disk with partitions: disk unpartition 0a.10.14 7. On both nodes, remove disk ownership: disk remove_ownership 8. On both nodes, verify that all drives connected to both nodes are unowned: disk show Managing disks using Data ONTAP | 37
  • 38. Example *> disk show Local System ID: 465245905 disk show: No disk match option show. 9. On both nodes, return to the LOADER menu: halt 10. On Node A only, begin zeroing the drives: a. At the LOADER prompt, boot Data ONTAP: boot_ontap b. Monitor the boot process, and when prompted, press Ctrl-C to display the boot menu. c. Select option 4, Clean configuration and initialize all disks. When the drives begin to be zeroed, a series of dots are printed to the console. 11. After the drives on Node A have been zeroing for a few minutes, repeat the previous step on Node B. The drives on both nodes that will be partitioned are zeroing . When the zeroing process is complete, the nodes return to the Node Setup wizard. 12. Restore your system configuration. 13. Confirm that the root aggregate on both nodes is composed of partitions: storage aggregate show-status The Position column shows shared and the usable size is a small fraction of the physical size: Owner Node: sys1 Aggregate: aggr0_1 (online, raid_dp) (block checksums) Plex: /aggr0_1/plex0 (online, normal, active, pool0) RAID Group /aggr0_1/plex0/rg0 (normal, block checksums) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- -------- shared 0a.10.10 0 BSAS 7200 73.89GB 828.0GB (normal) shared 0a.10.0 0 BSAS 7200 73.89GB 828.0GB (normal) shared 0a.10.11 0 BSAS 7200 73.89GB 828.0GB (normal) shared 0a.10.6 0 BSAS 7200 73.89GB 828.0GB (normal) shared 0a.10.5 0 BSAS 7200 73.89GB 828.0GB (normal) 14. Run System Setup to reconfigure the HA pair or rejoin the cluster, depending on your initial configuration. Related concepts Which drives are partitioned and used for the root aggregate on page 32 Requirements for using root-data partitioning on page 34 38 | Physical Storage Management Guide
  • 39. Related information Clustered Data ONTAP 8.3 Software Setup Guide Clustered Data ONTAP 8.3 High-Availability Configuration Guide Setting up an active-passive configuration on nodes using root-data partitioning When an HA pair is configured to use root-data partitioning by the factory, ownership of the data partitions is split between both nodes in the pair, for use in an active-active configuration. If you want to use the HA pair in an active-passive configuration, you must update partition ownership before creating your data aggregate. Before you begin • You should have decided which node will be the active node and which node will be the passive node. • Storage failover must be configured on the HA pair. About this task This task is performed on two nodes: Node A and Node B. All commands are input at the clustershell. This procedure is designed for nodes for which no data aggregate has been created from the partitioned disks. Steps 1. View the current ownership of the data partitions: storage aggregate show-spare-disks Example You can see that half of the data partitions are owned by one node and half are owned by the other node. All of the data partitions should be spare. cluster1::> storage aggregate show-spare-disks Original Owner: cluster1-01 Pool0 Partitioned Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size --------------------------- ----- ------ -------------- -------- -------- -------- 1.0.0 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.1 BSAS 7200 block 753.8GB 73.89GB 828.0GB 1.0.5 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.6 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.10 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.11 BSAS 7200 block 753.8GB 0B 828.0GB Managing disks using Data ONTAP | 39
  • 40. Original Owner: cluster1-02 Pool0 Partitioned Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size --------------------------- ----- ------ -------------- -------- -------- -------- 1.0.2 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.3 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.4 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.7 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.8 BSAS 7200 block 753.8GB 73.89GB 828.0GB 1.0.9 BSAS 7200 block 753.8GB 0B 828.0GB 12 entries were displayed. 2. Enter the advanced privilege level: set advanced 3. For each data partition owned by the node that will be the passive node, assign it to the active node: storage disk assign -force -data true -owner active_node_name -disk disk_name You do not need to include the partition as part of the disk name. Example You would enter a command similar to the following example for each data partition you need to reassign: storage disk assign -force -data true -owner cluster1-01 -disk 1.0.3 4. Confirm that all of the partitions are assigned to the active node. Example cluster1::*> storage aggregate show-spare-disks Original Owner: cluster1-01 Pool0 Partitioned Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size --------------------------- ----- ------ -------------- -------- -------- -------- 1.0.0 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.1 BSAS 7200 block 753.8GB 73.89GB 828.0GB 1.0.2 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.3 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.4 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.5 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.6 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.7 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.8 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.9 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.10 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.11 BSAS 7200 block 753.8GB 0B 828.0GB Original Owner: cluster1-02 Pool0 Partitioned Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size 40 | Physical Storage Management Guide
  • 41. --------------------------- ----- ------ -------------- -------- -------- -------- 1.0.8 BSAS 7200 block 0B 73.89GB 828.0GB 13 entries were displayed. Note that cluster1-02 still owns a spare root partition. 5. Return to administrative privilege: set admin 6. Create your data aggregate, leaving at least one data partition as spare: storage aggregate create new_aggr_name -diskcount number_of_partitions - node active_node_name The data aggregate is created and is owned by the active node. Related information Clustered Data ONTAP 8.3 High-Availability Configuration Guide Adding disks to a node You add disks to a node to increase the number of hot spares, to add space to an aggregate, or to replace disks. Before you begin You must have confirmed that your platform model supports the type of disk you want to add. About this task You use this procedure to add physical disks to your node. If you are administering a node that uses virtual disks, for example, a platform based on Data ONTAP-v technology, see the installation and administration guide that came with your Data ONTAP-v system. Steps 1. Check the NetApp Support Site for newer disk and shelf firmware and Disk Qualification Package files. If your node does not have the latest versions, you must update them before installing the new disk. 2. Install the disks according to the hardware guide for your disk shelf or the hardware and service guide for your platform. The new disks are not recognized until they are assigned to a node and pool. You can assign the new disks manually, or you can wait for Data ONTAP to automatically assign the new disks if your node follows the rules for disk autoassignment. Managing disks using Data ONTAP | 41
  • 42. 3. After the new disks have all been recognized, verify their addition and their ownership information: storage aggregate show-spare-disks You should see the new disks, owned by the correct node and in the correct pool. 4. Optional: Zero the newly added disks: storage disk zerospares Disks that have been used previously in a Data ONTAP aggregate must be zeroed before they can be added to another aggregate. Zeroing the disks now can prevent delays in case you need to quickly increase the size of an aggregate. The disk zeroing command runs in the background and can take hours to complete, depending on the size of the non-zeroed disks in the node. Result The new disks are ready to be added to an aggregate, used to replace an existing disk, or placed onto the list of hot spares. Related concepts How automatic ownership assignment works for disks on page 56 Guidelines for assigning ownership for disks on page 58 Related information NetApp Hardware Universe NetApp Downloads: Disk Drive and Firmware When you need to update the Disk Qualification Package The Disk Qualification Package (DQP) adds full support for newly qualified drives. Before you update drive firmware or add new drive types or sizes to a cluster, you must update the DQP. A best practice is to also update the DQP regularly; for example, every quarter or semi-annually. You need to download and install the DQP in the following situations: • Whenever you add a new drive type or size to the node For example, if you already have 1-TB drives and add 2-TB drives, you need to check for the latest DQP update. • Whenever you update the disk firmware • Whenever newer disk firmware or DQP files are available • Whenever you upgrade to a new version of Data ONTAP. The DQP is not updated as part of a Data ONTAP upgrade. 42 | Physical Storage Management Guide
  • 43. Related information NetApp Downloads: Disk Qualification Package NetApp Downloads: Disk Drive and Firmware Replacing disks that are currently being used in an aggregate You replace disks that are currently being used in an aggregate to swap out mismatched disks from a RAID group. Keeping your RAID groups homogeneous helps to optimize storage system performance. Before you begin You should already have an appropriate hot spare disk of the correct type, size, speed, and checksum type installed in your storage system. This spare disk must be assigned to the same system and pool as the disk it will replace. If the disk to be replaced is partitioned or in a storage pool, the spare disk (or its partitions, if applicable) can be owned by either node in the HA pair. For multi-disk carrier disks, you should have at least two hot spare disks available, to enable Data ONTAP to provide an optimal disk layout. About this task If you replace a smaller disk with a larger disk, the capacity of the larger disk is downsized to match that of the smaller disk; the usable capacity of the aggregate is not increased. Step 1. Replace the disk: storage disk replace -disk old_disk_name -replacement new_spare_disk_name -action start If you need to stop the disk replace operation, you can use the -action stop parameter. Result The old disk is converted to a spare disk, and the new disk is now used in the aggregate. Managing disks using Data ONTAP | 43
  • 44. Replacing a self-encrypting disk Replacing a self-encrypting disk (SED) is similar to replacing a regular disk, except that there are some extra steps you must take to reenable Storage Encryption after you replace the disk. Before you begin You should know the key used by the SEDs on your storage system so that you can configure the replacement SED to use the same key. Steps 1. Ensure that reconstruction has started by entering the following command: aggr status -r The status of the disk should display as "Reconstructing". 2. Remove the failed disk and replace it with a new SED, following the instructions in the hardware guide for your disk shelf model. 3. Assign ownership of the newly replaced SED by entering the following command: disk assign disk_name 4. Confirm that the new disk has been properly assigned by entering the following command: disk encrypt show You should see the newly added disk in the output. 5. Encrypt the disk by entering the following command: disk encrypt rekey key_id disk_name 6. Finalize the replacement process by entering the following command: disk encrypt lock disk_name The newly replaced SED is ready for use, and Storage Encryption is enabled and working on this system. Converting a data disk to a hot spare Data disks can be converted to hot spares by destroying the aggregate that contains them. Before you begin The aggregate to be destroyed cannot contain volumes. 44 | Physical Storage Management Guide
  • 45. About this task Converting a data disk to a hot spare does not change the ownership information for that disk. You must remove ownership information from a disk before moving it to another storage system. Step 1. Destroy the aggregate that contains the disk by entering the following command: storage aggregate delete -aggregate aggr_name All disks in use by that aggregate are converted to hot spare disks. Removing disks from a node How you remove a disk from a node depends how the disk is being used. By using the correct procedure, you can prevent unwanted AutoSupport notifications from being generated and ensure that the disk functions correctly if it is reused in another node. You cannot reduce the number of disks in an aggregate by removing data disks. The only way to reduce the number of data disks in an aggregate is to copy the data and transfer it to a new aggregate that has fewer data disks. If you are removing a partitioned drive, you must account for how both partitions are being used, and unpartition the drive before you remove it. If you are removing a disk in a multi-disk carrier, you must take special precautions. Related concepts Understanding root-data partitioning on page 30 How to determine when it is safe to remove a multi-disk carrier on page 28 Removing a failed disk A disk that is completely failed is no longer counted by Data ONTAP as a usable disk, and you can immediately disconnect the disk from the disk shelf. However, you should leave a partially failed disk connected long enough for the Rapid RAID Recovery process to complete. About this task If you are removing a disk because it has failed or because it is producing excessive error messages, you should not use the disk again in this or any other storage system. Steps 1. Find the disk ID of the failed disk by entering the following command: storage disk show -broken Managing disks using Data ONTAP | 45
  • 46. If the disk does not appear in the list of failed disks, it might be partially failed, with a Rapid RAID Recovery in process. In this case, you should wait until the disk is present in the list of failed disks (which means that the Rapid RAID Recovery process is complete) before removing the disk. 2. Determine the physical location of the disk you want to remove by entering the following command: storage disk set-led -disk disk_name 2 The fault LED on the face of the disk is lit for 2 minutes. 3. Remove the disk from the disk shelf, following the instructions in the hardware guide for your disk shelf model. Removing a hot spare disk Removing a hot spare disk requires you to remove ownership information from the disk. This prevents the disk from causing problems when it is inserted into another node, and notifies Data ONTAP that you are removing the disk, which prevents unwanted AutoSupport messages. About this task Removing a hot spare disk does not make the contents of that disk inaccessible. If you need absolute assurance that the data contained by this disk is irretrievable, you should sanitize the disk instead of completing this procedure. Steps 1. Find the disk name of the hot spare disk you want to remove: storage aggregate show-spare-disks 2. Determine the physical location of the disk you want to remove: storage disk set-led -disk disk_name The fault LED on the face of the disk is lit. 3. If disk ownership automatic assignment is on, turn it off: storage disk option modify -node node_name -autoassign off 4. If disk ownership automatic assignment is on for the node's HA partner, turn it off on the partner node also. 5. Remove the software ownership information from the disk: storage disk removeowner disk_name 6. Remove the disk from the disk shelf, following the instructions in the hardware guide for your disk shelf model. 46 | Physical Storage Management Guide
  • 47. 7. If you turned off disk ownership automatic assignment previously, turn it on now: storage disk option modify -node node_name -autoassign on 8. If you turned off disk ownership automatic assignment previously on the node's HA partner, turn it on for the partner node also. Related concepts How to determine when it is safe to remove a multi-disk carrier on page 28 Related tasks Using disk sanitization to remove data from disks on page 48 Removing a data disk The only time that you should remove a data disk from a storage system is if the disk is not functioning correctly. If you want to remove a data disk so that it can be used in another system, you must convert it to a hot spare disk first. About this task You can cause Data ONTAP to fail the disk immediately or allow a disk copy to finish before the disk is failed. If you do not fail the disk immediately, you must wait for the disk copy to finish before physically removing the disk. This operation might take several hours, depending on the size of the disk and the load on the storage system. Do not immediately fail a disk unless it is causing immediate performance or availability issues for your storage system. Depending on your storage system configuration, additional disk failures could result in data loss. Steps 1. Determine the name of the disk you want to remove. If the disk is reporting errors, you can find the disk name in the log messages that report disk errors. The name is prefixed with the word “Disk”. 2. Determine the physical location of the disk you want to remove by entering the following command: storage disk set-led -disk disk_name 2 The red LED on the face of the disk is lit for 2 minutes. 3. Take the appropriate action based on whether you need to fail the disk immediately. Managing disks using Data ONTAP | 47
  • 48. If you... Then... Can wait for the copy operation to finish (recommended) Enter the following command to pre-fail the disk: storage disk fail disk_name Data ONTAP pre-fails the specified disk and attempts to create a replacement disk by copying the contents of the pre-failed disk to a spare disk. If the copy operation is successful, then Data ONTAP fails the disk and the new replacement disk takes its place. If the copy operation fails, the pre-failed disk fails and the storage system operates in degraded mode until the RAID system reconstructs a replacement disk. Need to remove the disk immediately Enter the following command to cause the disk to fail immediately: storage disk fail -disk disk_name -immediate The disk fails and the storage system operates in degraded mode until the RAID system reconstructs a replacement disk. 4. Ensure that the disk you want to remove is shown as failed by entering the following command, and looking for its disk name: storage disk show -broken Do not remove the disk until its name appears in the list of failed disks. 5. Remove the failed disk from the disk shelf, following the instructions in the hardware guide for your disk shelf model. Related concepts About degraded mode on page 119 How to determine when it is safe to remove a multi-disk carrier on page 28 Using disk sanitization to remove data from disks Disk sanitization enables you to remove data from a disk or set of disks so that the data can never be recovered. Before you begin • The disks must be spare disks; they must be owned by a node, but not used in an aggregate. (If the disk is partitioned, neither partition can be in use in an aggregate.) • The disks cannot be part of a storage pool. 48 | Physical Storage Management Guide
  • 49. About this task When disk sanitization is enabled, it disables some Data ONTAP commands. After disk sanitization is enabled on a node, it cannot be disabled. If you need to remove data from disks using Storage Encryption, do not use this procedure. Use the procedure for destroying data on disks using Storage Encryption. Steps 1. Enter the nodeshell for the node that owns the disks you want to sanitize: system node run -node node_name 2. Enable disk sanitization: options licensed_feature.disk_sanitization.enable on You are asked to confirm the command because it is irreversible. 3. If the disks you want to sanitize are partitioned, unpartition each disk: disk unpartition disk_name 4. Sanitize the specified disks: disk sanitize start [-p pattern1|-r [-p pattern2|-r [-p pattern3|-r]]] [-c cycle_count] disk_list Attention: Do not turn off power to the node, disrupt the storage connectivity, or remove target disks while sanitizing. If sanitizing is interrupted during the formatting phase, the formatting phase must be restarted and allowed to finish before the disks are sanitized and ready to be returned to the spare pool. If you need to abort the sanitization process, you can do so by using the disk sanitize abort command. If the specified disks are undergoing the formatting phase of sanitization, the abort does not occur until the phase is complete. -p pattern1 -p pattern2 -p pattern3 specifies a cycle of one to three user-defined hex byte overwrite patterns that can be applied in succession to the disks being sanitized. The default pattern is three passes, using 0x55 for the first pass, 0xaa for the second pass, and 0x3c for the third pass. -r replaces a patterned overwrite with a random overwrite for any or all of the passes. -c cycle_count specifies the number of times that the specified overwrite patterns are applied. The default value is one cycle. The maximum value is seven cycles. disk_list specifies a space-separated list of the IDs of the spare disks to be sanitized. 5. If you want to check the status of the disk sanitization process: disk sanitize status [disk_list] Managing disks using Data ONTAP | 49
  • 50. 6. After the sanitization process is complete, return the disks to spare status by entering the following command for each disk: disk sanitize release disk_name 7. Return to the clustered Data ONTAP CLI: exit 8. Determine whether all of the disks were returned to spare status: storage aggregate show-spare-disks If... Then... All of the sanitized disks are listed as spares You are done. The disks are sanitized and in spare status. Some of the sanitized disks are not listed as spares Complete the following steps: a. Enter advanced privilege mode: set -privilege advanced b. Assign the unassigned sanitized disks to the appropriate node by entering the following command for each disk: storage disk assign -disk disk_name -owner node_name c. Return the disks to spare status by entering the following command for each disk: storage disk unfail -disk disk_name -s -q d. Return to administrative mode: set -privilege admin Result The specified disks are sanitized and designated as hot spares. The serial numbers of the sanitized disks are written to /etc/log/sanitized_disks. Related concepts How disk sanitization works on page 19 50 | Physical Storage Management Guide
  • 51. Stopping disk sanitization You can use the disk sanitize abort command to stop an ongoing sanitization process on one or more specified disks. Step 1. Enter the following command: disk sanitize abort disk_list This command is available through the nodeshell. If the specified disks are undergoing the disk formatting phase of sanitization, the process does not stop until the disk formatting is complete. Data ONTAP displays the message Sanitization abort initiated. After the process stops, Data ONTAP displays another message for each disk to inform you that sanitization is no longer in progress. Commands for managing disks You use the storage disk and storage aggregate commands to manage your disks. If you want to... Use this command... Display a list of spare disks, including partitioned disks, by owner storage aggregate show-spare-disks Display the disk RAID type, current usage, and RAID group by aggregate storage aggregate show-status Display the RAID type, current usage, aggregate, and RAID group, including spares, for physical disks storage disk show -raid Display a list of failed disks storage disk show -broken Illuminate the LED for a particular disk or shelf storage disk set-led Display the checksum type for a specific disk storage disk show -fields checksum- compatibility Display the checksum type for all spare disks storage disk show -fields checksum- compatibility -container-type spare Managing disks using Data ONTAP | 51
  • 52. If you want to... Use this command... Display disk connectivity and placement information storage disk show -fields disk,primary-port,secondary- name,secondary-port,shelf,bay Display pre-cluster disk names for specific disks storage disk show -disk -fields diskpathnames Display a list of disks in the maintenance center storage disk show -maintenance Unpartition a disk system node run -node local - command disk unpartition Zero all non-zeroed disks storage disk zerospares Related information Clustered Data ONTAP 8.3 Commands: Manual Page Reference Commands for displaying information about storage shelves You use the storage shelf show command to display configuration and error information for your disk shelves. If you want to display... Use this command... General information about shelf configuration and hardware status storage shelf show Detailed information for a specific shelf, including stack ID storage shelf show -shelf Unresolved, customer actionable, errors by shelf storage shelf show -errors Bay information storage shelf show -bay Connectivity information storage shelf show -connectivity Cooling information, including temperature sensors and cooling fans storage shelf show -cooling Information about I/O modules storage shelf show -module Port information storage shelf show -port 52 | Physical Storage Management Guide
  • 53. If you want to display... Use this command... Power information, including PSUs (power supply units), current sensors, and voltage sensors storage shelf show -power Related information Clustered Data ONTAP 8.3 Commands: Manual Page Reference Commands for displaying space usage information You use the storage aggregate and volume commands to see how space is being used in your aggregates and volumes and their Snapshot copies. To display information about... Use this command... Aggregates, including details about used and available space percentages, Snapshot reserve size, and other space usage information storage aggregate show storage aggregate show-space -snap- size-total,-used-including- snapshot-reserve How disks and RAID groups are used in an aggregate and RAID status storage aggregate show-status The amount of disk space that would be reclaimed if you deleted a specific Snapshot copy volume snapshot compute-reclaimable (advanced) The amount of space used by a volume volume show -fields size,used,available,percent-used volume show-space The amount of space used by a volume in the containing aggregate volume show-footprint Related information Clustered Data ONTAP 8.3 Commands: Manual Page Reference Managing disks using Data ONTAP | 53
  • 54. Managing ownership for disks Disk ownership determines which node owns a disk and what pool a disk is associated with. Data ONTAP stores ownership information directly on the disk. Types of disk ownership The HA or Disaster Recovery (DR) state of the system that owns a disk can affect which system has access to the disk. This means that there are several types of ownership for disks. Disk ownership information is set either by Data ONTAP or by the administrator, and recorded on the disk, in the form of the controller module's unique system ID (obtained from a node's NVRAM card or NVMEM board). Disk ownership information displayed by Data ONTAP can take one or more of the following forms. Note that the names used vary slightly depending on the context. • Owner (or Current owner) This is the system that can currently access the disk. • Original owner (or Home owner) If the system is in HA takeover, then Owner is changed to the system that took over the node, and Original owner or Home owner reflects the system that owned the disk before the takeover. • DR home owner If the system is in a MetroCluster switchover, DR home owner reflects the value of the Home owner field before the switchover occurred. Reasons to assign ownership of disks and array LUNs Storage system ownership must be assigned for disks and array LUNs before they become an effective part of your system. You must explicitly assign ownership for array LUNs. Disks can be automatically or manually assigned. You assign ownership of a disk or array LUN to accomplish the following actions: • Associate the disk or array LUN with a specific storage system. For a stand-alone system, all disks and array LUNs are owned by that system. In an HA configuration, the disks and array LUNs can be owned by either system. • Enable the disk or array LUN to be used and managed by the system that owns it. Unowned disks cannot be used as spares and do not receive the automatic firmware updates that owned disks do. 54 | Physical Storage Management Guide
  • 55. • Associate the disk or array LUN with a specific SyncMirror pool (when SyncMirror is in use). If SyncMirror is not in use, all disks and array LUNs are in pool0. How disks and array LUNs become available for use When you add a disk or array LUN to a system running Data ONTAP, the disk or array LUN goes through several stages before it can be used by Data ONTAP to store data or parity information. The process for making a disk available for use differs slightly from the process for making an array LUN available for use. Both processes are shown in the following diagram: Add to aggregate (optional) Install a new disk on a disk shelf Storage array Create array LUNs Make array LUNs available to Data ONTAP System running Data ONTAP Data ONTAP Manual assignment of array LUNs to a system running Data ONTAP Automatic or manual assignment of a new disk to a system running Data ONTAP Spare disk or array LUN It is owned by the storage system, but it cannot be used yet. In-use disk or array LUN The disk or LUN is in use by the system that owns it. Unowned disk or array LUN The process for disks includes the following actions: 1. The administrator physically installs the disk into a disk shelf. Data ONTAP can see the disk, but the disk is still unowned. 2. If the system is configured to support disk autoassignment, Data ONTAP assigns ownership for the disk; otherwise, the administrator must assign ownership of the disk manually. The disk is now a spare disk. 3. The administrator or Data ONTAP adds the disk to an aggregate. The disk is now in use by that aggregate. It could contain data or parity information. Managing ownership for disks | 55
  • 56. The process for array LUNs includes the following actions: 1. The storage array administrator creates the array LUN and makes it available to Data ONTAP. Data ONTAP can see the array LUN, but the array LUN is still unowned. 2. The Data ONTAP administrator assigns ownership of the array LUN to a Data ONTAP system. The array LUN is now a spare array LUN. 3. The Data ONTAP administrator adds the array LUN to an aggregate. The array LUN is now in use by that aggregate and is storing data. How automatic ownership assignment works for disks If your configuration follows some basic rules to avoid ambiguity, Data ONTAP can automatically assign ownership and pool membership for disks. Automatic disk autoassignment uses the configured autoassignment policy to determine the correct ownership for disks. The autoassignment policy can be stack, shelf, bay, or default. In most cases, all disks in a stack (or loop) can be assigned to the same node and pool. For mid- and high-end platforms, the default autoassignment policy is equivalent to the stack policy. For entry-level platforms (FAS2xxx), the default autoassignment policy is equivalent to the bay policy, which supports two controllers sharing the same shelf. For this policy, disks in odd bay numbers are assigned together, and disks in even bay numbers are assigned together. For partitioned disks, autoassignment operates on the container disk and also both of the partitions. If you decide to change the way Data ONTAP has assigned the disks, you can do so at any time. You can also change autoassignment to use a different assignment policy, although doing so does not affect currently assigned disks. If you need to temporarily remove disk ownership for one or more disks while you perform an administrative task, you must disable automatic ownership assignment first to prevent Data ONTAP from immediately reassigning ownership for those disks. Automatic ownership assignment is not available for array LUNs or virtual disks. Which disk autoassignment policy to use Usually, you can use the default autoassignment policy, which is equivalent to the stack policy for most systems, and to the bay policy for entry level systems (FAS2xxx). However, for some configurations, you might need to change the autoassignment policy. Use the appropriate autoassignment, based on your configuration: If you are using... Then use the autoassignment policy value of... Stand-alone entry level system stack 56 | Physical Storage Management Guide
  • 57. If you are using... Then use the autoassignment policy value of... Entry level systems in an HA configuration with a single, shared shelf bay Entry level systems in an HA configuration with one stack of two or more shelves shelf MetroCluster configurations with one stack per node, two or more shelves shelf All other configurations stack When automatic ownership assignment is invoked Automatic disk ownership assignment does not happen immediately after disks are introduced into the storage system. Automatic ownership assignment is invoked at the following times: • Every five minutes during normal system operation • Ten minutes after the initial system initialization This delay enables the person configuring the system enough time to finish the initial disk assignments so that the results of the automatic ownership assignment are correct. • Whenever you enable automatic ownership assignment. How disk ownership works for platforms based on Data ONTAP-v technology You manage ownership for virtual disks by using the same commands you use for physical disks. However, automatic ownership assignment works differently for virtual disks. Storage systems based on Data ONTAP-v technology, for example, Data ONTAP Edge systems, automatically assign ownership for all the virtual data disks you defined during the initial system setup. Automatic assignment is not available for any virtual disks you add after the initial system setup. For information about adding virtual disks and managing a storage system based on Data ONTAP-v technology, see the Data ONTAP Edge Installation and Administration Guide. Managing ownership for disks | 57
  • 58. Guidelines for assigning ownership for disks When you assign ownership for disks, you need to follow certain guidelines to maximize fault isolation and to keep automatic ownership assignment working. The guidelines are impacted by your autoassignment policy. Use these guidelines when you assign ownership for disks: • Assign disks so that your autoassignment policy group is homogenous. For example, if your autoassignment policy is stack (the default for most platforms), assign all disks on the same stack or loop to the same node and pool. However, if your autoassignment policy is bay (the default for FAS2xxx models), keep your disks in even-numbered bays and odd- numbered bays homogenous by node and pool. • If your autoassignment policy is stack, assign all stacks connected to the same adapter to the same pool. • Assign disks in the same multi-disk carrier to the same node and pool. Assigning ownership for disks Disks must be owned by a node before they can be used in an aggregate. If your cluster is not configured to use automatic disk ownership assignment, you must assign ownership manually. About this task This procedure is only for disks that are not partitioned. You can use the wildcard character to assign more than one disk at once. If you are reassigning a spare disk that is already owned by a different node, you must use the - force option for the storage disk assign command. You cannot reassign a disk that is in use in an aggregate. Steps 1. Display all unowned disks: storage disk show -container-type unassigned 2. Assign each disk: storage disk assign -disk disk_name -owner owner_name 58 | Physical Storage Management Guide
  • 59. Related concepts How automatic ownership assignment works for disks on page 56 Related tasks Assigning ownership for disks partitioned for root-data partitioning on page 59 Assigning ownership for disks partitioned for root-data partitioning For partitioned disks, there are three different entities for ownership: the container disk, the data partition, and the root partition. You can set the ownership of the container disk or the partitions manually or by using autoassignment—just as you do for unpartitioned disks. About this task You can think of the three owned entities (the container disk and the two partitions) as being collectively owned by the HA pair. The container disk and the two partitions do not all need to be owned by the same node in the HA pair as long as they are all owned by one of the nodes in the HA pair. However, when you use a partition in an aggregate, it must be owned by the same node that owns the aggregate. Steps 1. Display the current ownership for the partitioned disk: storage disk show -disk disk_name -partition-ownership 2. Enter the appropriate command, depending on which ownership entity you want to assign ownership for: If you want to assign ownership for the... Use this command... Container disk storage disk assign -disk disk_name -owner owner_name Data partition storage disk assign -disk disk_name -owner owner_name -data true Root partition storage disk assign -disk disk_name -owner owner_name -root true If any of the ownership entities are already owned, then you must include the -force option. Related concepts How automatic ownership assignment works for disks on page 56 Managing ownership for disks | 59
  • 60. Related tasks Assigning ownership for disks on page 58 Removing ownership from a disk Data ONTAP writes disk ownership information to the disk. Before you remove a spare disk or its shelf from a node, you should remove its ownership information so that it can be properly integrated into another node. Before you begin The disk you want to remove ownership from must meet the following requirements: • It must be a spare disk. You cannot remove ownership from a disk that is being used in an aggregate. • It cannot be in the maintenance center. • It cannot be undergoing sanitization. • It cannot be failed. It is not necessary to remove ownership from a failed disk. About this task If you have automatic disk assignment enabled, Data ONTAP could automatically reassign ownership before you remove the disk from the node. For this reason, you disable automatic ownership assignment until the disk is removed, and then reenable it. Steps 1. If disk ownership automatic assignment is on, turn it off: storage disk option modify -node node_name -autoassign off 2. If needed, repeat the previous step for the node's HA partner. 3. Remove the software ownership information from the disk: storage disk removeowner disk_name To remove ownership information from multiple disks, use a comma-separated list. Example storage disk removeowner sys1:0a.23,sys1:0a.24,sys1:0a.25 4. If the disk is partitioned for root-data partitioning, remove ownership from the partitions by entering both of the following commands: 60 | Physical Storage Management Guide
  • 61. storage disk removeowner disk_name -root true storage disk removeowner disk_name -data true Both partitions are no longer owned by any node. 5. If you turned off disk ownership automatic assignment previously, turn it on after the disk has been removed or reassigned: storage disk option modify -node node_name -autoassign on 6. If needed, repeat the previous step for the node's HA partner. Configuring automatic ownership assignment of disks You can configure Data ONTAP to automatically assign disk ownership according to a disk's stack, shelf, or bay. Before you begin • Your system must adhere to the requirements for automatic disk ownership. • If you have multiple stacks or shelves that must have different ownership, one disk must have been manually assigned on each stack or shelf so that automatic ownership assignment will work on each stack or shelf. About this task The behavior of the default autoassignment policy depends on the system model. For entry level models, the default policy is equivalent to the bay policy. For all other systems, it is equivalent to the stack policy. Steps 1. The action you take depends on whether you want to set up automatic ownership assignment at the stack (or loop), shelf, or bay level: If you want to... Then use the following command... Configure automatic ownership assignment at the stack or loop level storage disk option modify -autoassign-policy stack Configure automatic ownership assignment at the shelf level storage disk option modify -autoassign-policy shelf Managing ownership for disks | 61
  • 62. If you want to... Then use the following command... Configure automatic ownership assignment at the bay level storage disk option modify -autoassign-policy bay Turn off automatic ownership assignment storage disk option modify -autoassign off 2. Verify the automatic assignment settings for the disks: storage disk option show Example cluster1::> storage disk option show Node BKg. FW. Upd. Auto Copy Auto Assign Auto Assign Policy ------------- ------------- ------------ ------------- ------------------ cluster1-1 on on on default cluster1-2 on on on default Related concepts How automatic ownership assignment works for disks on page 56 Which disk autoassignment policy to use on page 56 When automatic ownership assignment is invoked on page 57 How you use the wildcard character with the disk ownership commands You can use the wildcard character (*) with some commands, including commands to manage disk ownership. However, you should understand how Data ONTAP expands the wildcard character to ensure that you operate on the correct set of disks. You can use the wildcard character with the following disk ownership commands: • storage disk assign • storage disk show • storage disk removeowner When you use the wildcard character with these commands, Data ONTAP expands it with zero or more characters to create a list of disk names that will be operated on by the command. This can be 62 | Physical Storage Management Guide
  • 63. very useful, for example, when you want to assign all of the disks attached to a particular port or switch. Note: Be careful when you use the wildcard character. It is accepted anywhere in the disk name string, and is a simple string substitution. Therefore, you might get unexpected results. For example, to operate on all disks on shelf 1 of stack 1, you would use the following syntax: 1.1.* However, if you left off the second “.”, as in the following snippet, you would operate on all disks in stacks 1, 10, 11, and so on: 1.1* Managing ownership for disks | 63
  • 64. Managing array LUNs using Data ONTAP For Data ONTAP to be able to use storage on a storage array, some tasks must be done on the storage array and some tasks must be done in Data ONTAP. For example, the storage array administrator must create array LUNs for Data ONTAP use and map them to Data ONTAP. You can then assign them to nodes running Data ONTAP. If the storage array administrator wants to make configuration changes to an array LUN after it is assigned to a node, for example to resize it, you might need to perform some activities in Data ONTAP before it is possible to reconfigure the LUN on the storage array. Related concepts How ownership for disks and array LUNs works on page 66 Data ONTAP systems that can use array LUNs on storage arrays V-Series (“V”) systems and new FAS platforms released in Data ONTAP 8.2.1 and later can use array LUNs if the proper license is installed. In discussions in the Data ONTAP and FlexArray Virtualization documentation, these systems are collectively referred to as Data ONTAP systems when it is necessary to make it clear which information applies to them and what information applies to storage arrays. Note: Starting with Data ONTAP 8.2.1, the capability of using LUNs on a storage array, formerly identified as V-Series functionality, has a new name—Data ONTAP FlexArray Virtualization Software. The capability of using array LUNs continues to be available as a licensed feature in Data ONTAP. Systems prior to Data ONTAP 8.2.1 that can use array LUNs The only systems released prior to Data ONTAP 8.2.1 that can use array LUNs are V-Series systems —systems with a “V” or “GF” prefix. A V-Series system is an open storage controller that virtualizes storage from storage array vendors, native disks, or both into a single heterogeneous storage pool. Note: Almost all Data ONTAP platforms released prior to Data ONTAP 8.2.1 were released with FAS and V-Series equivalent models (for example, a FAS6280 and a V6280 ). (For a few systems, there were no “V” equivalent models.) Although both types of models could access native disks, only the V-Series systems (a “V” or “GF” prefix) could attach to storage arrays. 64 | Physical Storage Management Guide
  • 65. Systems in Data ONTAP 8.2.1 and later that can use array LUNs Starting with Data ONTAP 8.2.1, the model for how platforms are released and the storage they can use changes. Attaching to storage arrays is no longer limited to V-Series systems. Starting with Data ONTAP 8.2.1, all new platforms are released as a single hardware model. This single hardware model has a FAS prefix; there are no longer separate “V” and FAS models for new platforms. If the V_StorageAttach license package is installed on a new FAS model, it can attach to storage arrays. (This is the same license required on a V-Series system.) Important: FAS systems released prior to Data ONTAP 8.2.1 cannot use LUNs on storage arrays, even if they are upgraded to Data ONTAP 8.2.1 or later; only the “V” equivalent of a platform can use array LUNs. Installing the license for using array LUNs The V_StorageAttach license must be installed on each Data ONTAP node that you want to use with array LUNs. It is not a single license for the cluster. Array LUNs cannot be used in aggregates until a license is installed. Before you begin • The cluster must be installed. • You must have the license key for the V_StorageAttach license. NetApp Support About this task You need not perform this procedure if the license key for the V_StorageAttach package is already installed. If the Data ONTAP system is ordered with disks, the factory typically installs the license package for you. Alternatively, many customers install all necessary licenses early in the installation process. Steps 1. For each Data ONTAP node in the cluster for use with array LUNs, enter the following command on the node: system license add license key Managing array LUNs using Data ONTAP | 65
  • 66. Example vgv3170f41a> license Serial Number: nnnnnnnn Owner: mysystem1a Package Type Description Expiration ----------------- ------- --------------------- -------------------- V_StorageAttach license Virtual Attached Storage 2. Look at the output to confirm that the V_StorageAttach package is shown. How ownership for disks and array LUNs works Disk and array LUN ownership determines which node owns a disk or array LUN and what pool a disk or array LUN is associated with. Understanding how ownership works enables you to maximize storage redundancy and manage your hot spares effectively. Data ONTAP stores ownership information directly on the disk or array LUN. Reasons to assign ownership of disks and array LUNs Storage system ownership must be assigned for disks and array LUNs before they become an effective part of your system. You must explicitly assign ownership for array LUNs. Disks can be automatically or manually assigned. You assign ownership of a disk or array LUN to accomplish the following actions: • Associate the disk or array LUN with a specific storage system. For a stand-alone system, all disks and array LUNs are owned by that system. In an HA configuration, the disks and array LUNs can be owned by either system. • Enable the disk or array LUN to be used and managed by the system that owns it. Unowned disks cannot be used as spares and do not receive the automatic firmware updates that owned disks do. • Associate the disk or array LUN with a specific SyncMirror pool (when SyncMirror is in use). If SyncMirror is not in use, all disks and array LUNs are in pool0. How disks and array LUNs become available for use When you add a disk or array LUN to a system running Data ONTAP, the disk or array LUN goes through several stages before it can be used by Data ONTAP to store data or parity information. The process for making a disk available for use differs slightly from the process for making an array LUN available for use. Both processes are shown in the following diagram: 66 | Physical Storage Management Guide
  • 67. Add to aggregate (optional) Install a new disk on a disk shelf Storage array Create array LUNs Make array LUNs available to Data ONTAP System running Data ONTAP Data ONTAP Manual assignment of array LUNs to a system running Data ONTAP Automatic or manual assignment of a new disk to a system running Data ONTAP Spare disk or array LUN It is owned by the storage system, but it cannot be used yet. In-use disk or array LUN The disk or LUN is in use by the system that owns it. Unowned disk or array LUN The process for disks includes the following actions: 1. The administrator physically installs the disk into a disk shelf. Data ONTAP can see the disk, but the disk is still unowned. 2. If the system is configured to support disk autoassignment, Data ONTAP assigns ownership for the disk; otherwise, the administrator must assign ownership of the disk manually. The disk is now a spare disk. 3. The administrator or Data ONTAP adds the disk to an aggregate. The disk is now in use by that aggregate. It could contain data or parity information. The process for array LUNs includes the following actions: 1. The storage array administrator creates the array LUN and makes it available to Data ONTAP. Data ONTAP can see the array LUN, but the array LUN is still unowned. 2. The Data ONTAP administrator assigns ownership of the array LUN to a Data ONTAP system. The array LUN is now a spare array LUN. 3. The Data ONTAP administrator adds the array LUN to an aggregate. The array LUN is now in use by that aggregate and is storing data. Managing array LUNs using Data ONTAP | 67
  • 68. What it means for Data ONTAP to own an array LUN Data ONTAP cannot use an array LUN presented to it by a storage array until you configure a logical relationship in Data ONTAP that identifies a specific system running Data ONTAP as the owner of the array LUN. A storage array administrator creates array LUNs and makes them available to specified FC initiator ports of storage systems running Data ONTAP. (The process for how to do this varies among storage array vendors.) When you assign an array LUN to a system running Data ONTAP, Data ONTAP writes data to the array LUN to identify that system as the owner of the array LUN. Thereafter, Data ONTAP ensures that only the owner can write data to and read data from the array LUN. From the perspective of Data ONTAP, this logical relationship is referred to as disk ownership because Data ONTAP considers an array LUN to be a virtual disk. From the perspective of Data ONTAP, you are assigning disks to a storage system. An advantage of the disk ownership scheme is that you can make changes through the Data ONTAP software that, on typical hosts, must be done by reconfiguring hardware or LUN access controls. For example, through Data ONTAP you can balance the load of requests among a group of systems running Data ONTAP by moving data service from one system to another, and the process is transparent to most users. You do not need to reconfigure hardware or the LUN access controls on the storage array to change which system running Data ONTAP is the owner and, therefore, servicing data requests. Attention: The Data ONTAP software-based scheme provides ownership control only for storage systems running Data ONTAP; it does not prevent a different type of host from overwriting data in an array LUN owned by a system running Data ONTAP. Therefore, if multiple hosts are accessing array LUNs through the same storage array port, be sure to use LUN security on your storage array to prevent the systems from overwriting each other's array LUNs. Array LUN reconfiguration, such as resizing the array LUN, must be done from the storage array. Before such activities can occur, you must release Data ONTAP ownership of the array LUN. Why you might assign array LUN ownership after installation For a Data ONTAP system ordered with disk shelves, you are not required to set up the system to work with array LUNs during initial installation. In the case of a Data ONTAP system using only array LUNs, you need to assign only two array LUNs during initial installation. If you ordered your Data ONTAP system with disk shelves, you do not need to assign any array LUNs initially because the factory installs the root volume on a disk for you. If you are using only array LUNs, you must configure one array LUN for the root volume and one array LUN as a spare for core dumps during initial installation. In either case, you can assign ownership of additional array LUNs to your system at any time after initial installation. After you configure your system, you might assign ownership of an array LUN in the following circumstances: 68 | Physical Storage Management Guide
  • 69. • You ordered your Data ONTAP system with native disk shelves and did not set up your system to work with array LUNs initially • You left some LUNs that the storage array presented to Data ONTAP unowned, and you now need to use the storage • Another system released the ownership of a particular array LUN and you want this system to use the LUN • The storage array administrator has not made the LUNs available to Data ONTAP during the initial system configuration, and you want to use the storage. Examples showing when Data ONTAP can use array LUNs After an array LUN has been assigned to a storage system, it can be added to an aggregate and used for storage or it can remain a spare LUN until it is needed for storage. No storage system owns the LUNs yet In this example, the storage array administrator made the array LUNs available to Data ONTAP. However, system vs1 has not yet been configured to “own” any of the LUNs. Therefore, it cannot read data from or write data to any array LUNs on the storage array: Only some array LUNs are owned In this example, vs1 was configured to own array LUNs 1 and 2, but not array LUNs 3 and 4. LUNs 3 and 4 are still available to Data ONTAP, however, and can be assigned to a storage system later: Data ONTAP used LUN 1 for the root volume. System vs1 can read data from and write data to LUN 1, because LUN 1 is in an aggregate. LUN 2 remains a spare LUN because it has not yet been added to an aggregate. System vs1 cannot read data from and write data to LUN 2 while it is a spare. Managing array LUNs using Data ONTAP | 69
  • 70. After you perform initial setup of the storage system, you could configure vs1 to also own LUN 3, LUN 4, both, or neither, depending on your storage needs. Ownership of LUNs in an HA pair In this example, two storage systems running Data ONTAP are configured in an HA pair. In an HA pair, only one node can be the owner of a particular LUN, but both nodes must be able to see the same LUNs so that the partner can take over if the owning node becomes unavailable. LUN 1 through LUN 4 were created on the storage array and mapped to the ports on the storage array to which the storage systems are connected. All four LUNs are visible to each node in the HA pair. Assume that during initial setup vs1 was assigned ownership of LUN 1 and LUN 2. LUN 1 was automatically added to the root volume, so LUN 1 is now “in use” by vs1. LUN 2 remains a spare until it is explicitly added to an aggregate on vs1. Similarly, assume that during initial setup vs2 was assigned ownership of LUN 3 and LUN 4, with LUN 3 assigned to the root volume. LUN 4 remains a spare LUN until it is explicitly added to an aggregate. The key points of this example are as follows: 70 | Physical Storage Management Guide
  • 71. • By deploying the storage systems in an HA pair, one system can take over services for its partner if the partner becomes unavailable. • Only one storage system can own a specific array LUN. However, all array LUNs assigned to a node in an HA pair must be visible to—but not assigned to or owned by—the other node in the HA pair. • By deploying two switches, if one switch fails, the other switch provides the alternate path to the storage array. • Both switches must be zoned correctly so that each storage system in the HA pair can see the array LUNs owned by its partner. Assigning ownership of array LUNs Array LUNs must be owned by a node before they can be added to an aggregate to be used as storage. Before you begin • Back-end configuration testing (testing of the connectivity and configuration of devices behind the Data ONTAP systems) must be completed. • Array LUNs that you want to assign must be presented to the Data ONTAP systems. About this task You can assign ownership of array LUNs that have the following characteristics: • They are unowned. • They have no storage array configuration errors, such as the following: ◦ The array LUN is smaller than or larger than the size that Data ONTAP supports. ◦ The LDEV is mapped on only one port. ◦ The LDEV has inconsistent LUN IDs assigned to it. ◦ The LUN is available on only one path. Data ONTAP issues an error message if you try to assign ownership of an array LUN with back-end configuration errors that would interfere with the Data ONTAP system and the storage array operating together. You must fix such errors before you can proceed with array LUN assignment. Data ONTAP alerts you if you try to assign an array LUN with a redundancy error: for example, all paths to this array LUN are connected to the same controller or only one path to the array LUN. You can fix a redundancy error before or after assigning ownership of the LUN. Managing array LUNs using Data ONTAP | 71
  • 72. Steps 1. Enter the following command to see the array LUNs that have not yet been assigned to a node: storage disk show -container-type unassigned 2. Enter the following command to assign an array LUN to this node: storage disk assign -disk arrayLUNname -owner nodename If you want to fix a redundancy error after disk assignment instead of before, you must use the - force parameter with the storage disk assign command. Related concepts How ownership for disks and array LUNs works on page 66 Related tasks Modifying assignment of spare array LUNs on page 73 Verifying the back-end configuration It is important to detect and resolve any configuration errors before you bring online the Data ONTAP configuration with array LUNs in a production environment. You start installation verification by using storage array config show command. The storage array config show command shows how storage arrays connect to the cluster. If Data ONTAP detects an error in the back-end configuration, the following message is displayed at the bottom of the storage array config show output: Warning: Configuration errors were detected. Use 'storage errors show' for detailed information. You then use the storage errors show output to see details of the problem, at the LUN level. You must fix any errors shown by the storage errors show command. Related information FlexArray Virtualization Implementation Guide for Third-Party Storage 72 | Physical Storage Management Guide
  • 73. Modifying assignment of spare array LUNs You can change the ownership of a spare array LUN to another node. You might want to do this for load balancing over the nodes. Steps 1. At the console of the node that owns the array LUN you want to reassign, enter the following command to see a list of spare array LUNs on the node: storage disk show -owner local The array LUNs owned by the node, both spares and LUNs in aggregates, are listed. 2. Confirm that the LUN you want to reassign to another node is a spare LUN. 3. Enter the following command to assign ownership of the array LUN to another node: storage disk assign arrayLUNname -owner new_owner_name -force Note: The array LUN ownership is not changed if the -force option is not used or if the array LUN was already added to an aggregate. 4. Enter the following command to verify that the ownership of the spare array LUN was changed to the other node: storage disk show -owner local The spare array LUN that you changed to the new owner should no longer appear in the list of spares. If the array LUN still appears, repeat the command to change ownership. 5. On the destination node, enter the following command to verify that the spare array LUN whose ownership you changed is listed as a spare owned by the destination node: storage disk show -owner local After you finish You must add the array LUN to an aggregate before it can be used for storage. Related concepts How ownership for disks and array LUNs works on page 66 Related tasks Assigning ownership of array LUNs on page 71 Managing array LUNs using Data ONTAP | 73
  • 74. Array LUN name format In Data ONTAP 8.3, the name assigned to an array LUN has a new format to ensure that the name is unique across a cluster. The array LUN name consists of two components and looks like the following: <array_prefix>.<offset>, for example EMC-1.1. • The array_prefix is a unique prefix that Data ONTAP assigns by default to each storage array. This field is composed of <array_name-array_instance> (EMC-1 in this case). array_name can be denoted by the first three letters of the name of the vendor. If there is more than one array from the same vendor, the value of array_instance proceeds in an ascending order. • The offset is the ascending virtual disk number that Data ONTAP assigns to each LUN. It is independent of the LUN ID of the host. You can modify the <array_prefix> field by using the storage array modify -name - prefix command. Pre-cluster array LUN name format Before a node joins to a cluster or when the system is in the maintenance mode, the array LUN name follows a format used before Data ONTAP 8.3, the pre-cluster format. In this format, the array LUN name is a path-based name that includes the devices in the path between the Data ONTAP system and the storage array, ports used, and the SCSI LUN ID on the path that the storage array presents externally for mapping to hosts. On a Data ONTAP system that supports array LUNs, each array LUN can have multiple names because there are multiple paths to each LUN. 74 | Physical Storage Management Guide
  • 75. Array LUN name format for Data ONTAP systems Configuration Array LUN name format Component descriptions Direct-attached node- name.adapter.idlun-id node-name is the name of the clustered node. With clustered Data ONTAP, the node name is prepended to the LUN name so that the path-based name is unique within the cluster. adapter is the adapter number on the Data ONTAP system. id is the channel adapter port on the storage array. lun-id is the array LUN number that the storage array presents to hosts. Example: node1.0a.0L1 Fabric-attached node-name:switch- name:port.idlun-id node-name is the name of the node. With clustered Data ONTAP, the node name is prepended to the LUN name so that the path-based name is unique within the cluster. switch-name is the name of the switch. port is the switch port that is connected to the target port (the end point). id is the device ID. lun-id is the array LUN number that the storage array presents to hosts. Example: node1:brocade3:6.126L1 Related concepts Pre-cluster drive name formats on page 15 Managing array LUNs using Data ONTAP | 75
  • 76. Checking the checksum type of spare array LUNs If you plan to add a spare array LUN to an aggregate by specifying its name, you need to make sure that the checksum type of the array LUN you want to add is the same as the aggregate checksum type. About this task You cannot mix array LUNs of different checksum types in an array LUN aggregate. The checksum type of the aggregate and the checksum type of the array LUNs added to it must be the same. If you specify a number of spare array LUNs to be added to an aggregate, by default Data ONTAP selects array LUNs of the same checksum type as the aggregate. Note: The checksum type of all newly created aggregates using zoned checksum array LUNS is advanced zoned checksum (AZCS). Zoned checksum type continues to be supported for existing zoned aggregates. Zoned checksum spare array LUNs added to an existing zoned checksum aggregate continue to be zoned checksum array LUNs. Zoned checksum spare array LUNs added to an AZCS checksum type aggregate use the AZCS checksum scheme for managing checksums. Step 1. Check the checksum type of the spare array LUNs by entering the following command: storage disk show -fields checksum-compatibility -container-type spare You can add a block checksum array LUN to a block checksum aggregate and a zoned array LUN to an advanced zoned checksum (AZCS) aggregate. Related tasks Changing the checksum type of an array LUN on page 76 Changing the checksum type of an array LUN You must change the checksum type of an array LUN if you want to add it to an aggregate that has a different checksum type than the checksum type of the LUN. Before you begin You must have reviewed the tradeoffs between performance in certain types of workloads and storage capacity utilization of each checksum type. The FlexArray Virtualization Installation Requirements and Reference Guide contains information about checksum use for array LUNs. You can also contact your Sales Engineer for details about using checksums. 76 | Physical Storage Management Guide
  • 77. About this task • You must assign a zoned checksum type to an array LUN that you plan to add to an advanced zoned checksum (AZCS) aggregate. When a zoned checksum array LUN is added to an AZCS aggregate, it becomes an advanced zoned checksum array LUN. Similarly, when a zoned checksum array LUN is added to a zoned aggregate, it is a zoned checksum type. • You cannot modify the checksum of array LUNs while assigning ownership. You can modify the checksum only on already assigned array LUNs. Step 1. Enter the following command to change the checksum type: storage disk assign -disk disk name -owner owner -c new_checksum_type disk name is the array LUN whose checksum type you want to change. owner is the node to which the array LUN is assigned. new_checksum_type can be block or zoned. Example storage disk assign -disk EMC-1.1 -owner system147b -c block The checksum type of the array LUN is changed to the new checksum type you specified. Related tasks Checking the checksum type of spare array LUNs on page 76 Prerequisites to reconfiguring an array LUN on the storage array If an array LUN has already been assigned (through Data ONTAP) to a particular Data ONTAP system, the information that Data ONTAP wrote to the array LUN must be removed before the storage administrator attempts to reconfigure the array LUN on the storage array. When the storage array presents an array LUN to Data ONTAP, Data ONTAP collects information about the array LUN (for example, its size) and writes that information to the array LUN. Data ONTAP cannot dynamically update information that it wrote to an array LUN. Therefore, before the storage array administrator reconfigures an array LUN, you must use Data ONTAP to change the state of the array LUN to unused. The array LUN is unused from the perspective of Data ONTAP. While changing the state of the array LUN to unused, Data ONTAP does the following: • Terminates I/O operations to the array LUN Managing array LUNs using Data ONTAP | 77
  • 78. • Removes the label for RAID configuration information and the persistent reservations from the array LUN, which makes the array LUN unowned by any Data ONTAP system After this process finishes, no Data ONTAP information remains in the array LUN. You can do the following after the array LUN's state is changed to unused: • Remove the mapping of the array LUN to Data ONTAP and make the array LUN available to other hosts. • Resize the array LUN or change its composition. If you want Data ONTAP to resume using the array LUN after its size or composition is changed, you must present the array LUN to Data ONTAP again, and assign the array LUN to a Data ONTAP system again. Data ONTAP is aware of the new array LUN size or composition. Related tasks Changing array LUN size or composition on page 78 Changing array LUN size or composition Reconfiguring the size or composition of an array LUN must be done on the storage array. If an array LUN has already been assigned to a Data ONTAP system, you must use Data ONTAP to change the state of the array LUN to unused before the storage array administrator can reconfigure it. Before you begin The array LUN must be a spare array LUN before you can change its state to unused. Steps 1. On the Data ONTAP system, enter the following command to remove ownership information: storage disk removeowner -disk arrayLUNname 2. On the storage array, complete the following steps: a. Unmap (unpresent) the array LUN from the Data ONTAP systems so that they can no longer see the array LUN. b. Change the size or composition of the array LUN. c. If you want Data ONTAP to use the array LUN again, present the array LUN to the Data ONTAP systems again. At this point, the array LUN is visible to the FC initiator ports to which the array LUN was presented, but it cannot be used by any Data ONTAP systems yet. 78 | Physical Storage Management Guide
  • 79. 3. Enter the following command on the Data ONTAP system that you want to be the owner of the array LUN: storage disk assign -disk arrayLUNname -owner nodename After the ownership information is removed, the array LUN cannot be used by any Data ONTAP system until the array LUN is assigned again to a system. You can leave the array LUN as a spare or add it to an aggregate. You must add the array LUN to an aggregate before the array LUN can be used for storage. Related concepts Prerequisites to reconfiguring an array LUN on the storage array on page 77 Removing one array LUN from use by Data ONTAP If the storage array administrator no longer wants to use a particular array LUN for Data ONTAP, you must remove the information that Data ONTAP wrote to the LUN (for example, size and ownership) before the administrator can reconfigure the LUN for use by another host. Before you begin If the LUN that the storage array administrator no longer wants Data ONTAP to use is in an aggregate, you must take the aggregate offline and destroy the aggregate before starting this procedure. Taking an aggregate offline and destroying it changes the data LUN to a spare LUN. Step 1. Enter the following command: storage disk removeowner -disk LUN_name LUN_name is the name of the array LUN. Preparing array LUNs before removing a Data ONTAP system from service You must release the persistent reservations on all array LUNs assigned to a Data ONTAP system before removing the system from service. About this task When you assign Data ONTAP ownership of an array LUN, Data ONTAP places persistent reservations (ownership locks) on that array LUN to identify which Data ONTAP system owns the LUN. If you want the array LUNs to be available for use by other types of hosts, you must remove the persistent reservations that Data ONTAP put on those array LUNs; some arrays do not allow you Managing array LUNs using Data ONTAP | 79
  • 80. to destroy a reserved LUN if you do not remove the ownership and persistent reservations that Data ONTAP wrote to that LUN. For example, the Hitachi USP storage array does not have a user command for removing persistent reservations from LUNs. If you do not remove persistent reservations through Data ONTAP before removing the Data ONTAP system from service, you must call Hitachi technical support to remove the reservations. Contact technical support for instructions about how to remove persistent reservations from LUNs before removing a Data ONTAP system from service. 80 | Physical Storage Management Guide
  • 81. Securing data at rest with Storage Encryption You can protect your data at rest from unauthorized access with Storage Encryption. To use this feature, you must understand what Storage Encryption is and how it works, how to set it up and manage it, and how you can destroy data on disks using Storage Encryption. Introduction to Storage Encryption Overview of Storage Encryption concepts, functionality, benefits, and limitations. What Storage Encryption is Storage Encryption is an optional feature that you can enable for additional data protection. It is available on certain supported storage controllers and disk shelves that contain disks with built-in encryption functionality. In a standard storage environment, data is written to disk in cleartext format. This makes the data vulnerable to potential exposure to unauthorized users when disks removed from a storage system are lost or stolen. When you enable Storage Encryption, the storage system protects your data at rest by storing it on self-encrypting disks. The authentication keys used by the self-encrypting disks are stored securely on external key management servers. Purpose of the external key management server An external key management server is a third-party system in your storage environment that securely manages authentication keys used by the self-encrypting disks in the storage system. You link the external key management server to other systems that use authentication or encryption keys such as your storage system. The storage system uses a secure SSL connection to connect to the external key management server to store and retrieve authentication keys. The communication between the storage system and key management server uses the Key Management Interoperability Protocol (KMIP). The external key management server securely stores authentication or encryption keys entrusted to it and provides them upon demand to authorized linked systems. This provides an additional level of security by storing authentication keys separate from the storage system. Additionally, authentication keys are always handled and stored securely. The keys are never displayed in cleartext. You must link at least one key management server to the storage system during the Storage Encryption setup and configuration process. You should link multiple key management servers for redundancy. If the only key management server in the environment becomes unavailable, access to protected data might become unavailable until the key management server is available again. For 81
  • 82. example, when the storage system needs to unlock self-encrypting disks but cannot retrieve the authentication key from the key management server because it is unavailable. You can specify up to four key servers during or after setup for redundancy. For a list of supported key management servers, see the Interoperability Matrix. How Storage Encryption works Storage Encryption occurs at the firmware level of disks that are equipped with special firmware and hardware to provide the additional security, also known as self-encrypting disks (SEDs). SEDs can operate either in unprotected mode like regular disks, or in protected mode requiring authentication after the power-on process. SEDs always encrypt data for storage. In unprotected mode, the encryption key needed to decrypt and access the data is freely available. In protected mode, the encryption key is protected and requires authentication to be used. When you first enable and configure Storage Encryption on a storage system using SEDs, you create an authentication key that the storage system uses to authenticate itself to the SEDs. You configure the storage system with the IP address to one or more external key management servers that securely stores the authentication key. The storage system communicates with the key management servers at boot time to retrieve the authentication keys. Data ONTAP requires the authentication keys to authenticate itself to the SEDs any time after the SEDs are power-cycled. If the authentication is successful, the SEDs are unlocked. The SEDs use the authentication key to decrypt the data encryption keys stored inside the disk. When presented with a read request, SEDs automatically decrypt the stored data before passing it on to the storage system. When presented with a write request from the storage system, SEDs automatically encrypt the data before writing the data to the disk's storage platters. When the SEDs are locked, Data ONTAP must successfully authenticate itself to the disk before the SEDs allow data to be read or written. When locked, SEDs require authentication each time the disk is powered on. Encryption and decryption happens without a perceptible disk performance decrease or boot time increase. Storage Encryption does not require a separate license key. The only additional required component is an external key management server. When you halt and power down the storage system, including the disk shelves containing SEDs, the disks are locked again and the data becomes inaccessible. Disk operations with SEDs Most of the disk-related operations are identical for SEDs and regular disks. Because storage encryption happens at a very low level, specifically the disk firmware, it does not affect any higher level functionality. The storage controller sees SEDs the same as regular disks, and all functionality remains the same. There are some additional options and requirements with SEDs: 82 | Physical Storage Management Guide
  • 83. • Sanitizing disks There are additional options to sanitize disks when using SEDs. • Destroying disks An additional option enables you to make the disks permanently inaccessible. Benefits of using Storage Encryption There are several scenarios where using Storage Encryption provides significant benefits by protecting data from unauthorized access when disks removed from a storage system have fallen into the wrong hands. Data protection in case of disk loss or theft Storage Encryption protects your data if disks are lost or stolen. Someone who comes into possession of disks that store data using Storage Encryption cannot access the data. Without the authentication key that is required to authenticate and unlock the disks, all attempts to read or write data result in an error message returned by the SEDs. Circumventing the disk authentication by moving the platters into another disk without encryption firmware would be unsuccessful as well. The data stored on the platters appears as ciphertext and is fully protected from unauthorized access. Data protection when returning disks to vendors Storage Encryption protects your data when you return disks to vendors. The following three options are available to protect data on disks that are removed from a storage system and returned to a vendor: • If the SED is owned by a storage system, it requires authentication to access the data. Since the vendor does not know, or have access to, the authentication key, the vendor cannot access data on the disk. • If you sanitize the disk before returning it to a vendor, it changes the encryption key to a new unknown key. Any subsequent attempts to read data from the disk result in random data. • If you "destroy" the disk, it changes the encryption key to a random unknown key, it changes the authentication key to a random unknown key, and permanently locks the disk, preventing any further decryption of the data and access to the disk. Related tasks Sanitizing disks using Storage Encryption before return to vendor on page 102 Securing data at rest with Storage Encryption | 83
  • 84. Data protection when moving disks to end-of-life Storage Encryption protects your data when moving a disk to an end-of-life state. You can protect data on a disk by changing the authentication key to a random value that is not stored and permanently locking the drive. This prevents any further decryption of the data and access to the disk. Related tasks Setting the state of disks using Storage Encryption to end-of-life on page 103 Data protection through emergency data shredding Storage Encryption protects your data in emergency situations by allowing you to instantaneously prevent access to the data on the disk. This might include extreme scenarios where power to the storage system or the key management server (or both) is not available, or one or both have fallen into possession of a hostile third-party. Related tasks Emergency shredding of data on disks using Storage Encryption on page 104 Limitations of Storage Encryption You must keep certain limitations in mind when using Storage Encryption. • For the latest information about which storage systems, disk shelves, and key management servers are supported with Storage Encryption, see the Interoperability Matrix. • All disks in the storage system and optional attached disk shelves must have encryption functionality to be able to use Storage Encryption. You cannot mix regular non-encrypting disks with self-encrypting disks. • Storage Encryption key_manager commands are only available for local nodes. They are not available in takeover mode for partner nodes. • Do not configure Storage Encryption to use 10 Gigabit network interfaces for communication with key management servers. This limitation does not apply to serving data. • Storage Encryption supports a maximum of 128 authentication keys per key management server. You receive a warning when the number of stored authentication keys reaches 100. You cannot create new authentication keys when the number of stored authentication keys reaches the limit of 128. You must then delete unused authentication keys before you can create new ones. • Storage Encryption supports KMIP 1.0 and 1.1 for communication with key management servers. 84 | Physical Storage Management Guide
  • 85. Related information Interoperability Matrix Tool: mysupport.netapp.com/matrix Setting up Storage Encryption During initial setup, your storage system checks whether it is properly configured with self- encrypting disks and is running a version of Data ONTAP that supports Storage Encryption. If the check is successful, you can then launch the Storage Encryption setup wizard after completion of the storage system setup wizard. Information to collect before configuring Storage Encryption You must gather certain information to successfully set up Storage Encryption on your storage system. Information to collect Details Required Optional Network interface name You must provide the name of the network interface the storage system should use to communicate with external key management servers. Note: Do not configure 10-Gigabit network interfaces for communication with key management servers. x Network interface IP address You must provide the IP address of the network interface. This IP address can be in either IPv4 or IPv6 format. x IPv6 network prefix length For IPv6 addresses, you must provide the network prefix length. You can either provide it by appending a slash (/) and the network prefix length directly to the IPv6 address when entering it, or you can enter the network prefix length separately when prompted after entering the address. x Network interface subnet mask You must provide the subnet mask of the network interface. x Network interface gateway IP address You must provide the IP address for the network interface gateway. x Securing data at rest with Storage Encryption | 85
  • 86. Information to collect Details Required Optional IP addresses for external key management servers You must link the storage system to at least one external key management server during setup. You should add two or more external key management servers to prevent having a single point of failure. If you add only one external key management server and it fails, you can lose access to your data. If you specify IPv6 addresses for external key management servers, you must also provide an IPv6 address for the storage system network interface. x IP address for additional external key management servers You can link the storage system to multiple additional external key management servers during setup for redundancy. x Port number for each external key management server You must provide the port number that each key management server listens on. The port number must be the same for all key management servers. x Public SSL certificate for storage system You must provide a public SSL certificate for the storage system to link it to the external key management server. x Private SSL certificate for storage system You must provide a private SSL certificate for the storage system. x Public SSL certificate for external key management servers You must provide a public SSL certificate for each external key management server to link it to the storage system. x Key tag name You can provide a name that is used to identify all keys belonging to a particular storage system. The default key tag name is the system's host name. x Using SSL for secure key management communication The storage system and key management servers use SSL connections to keep the communication between them secure. This requires you to obtain and install various SSL certificates for the storage system and each key management server before you can set up and configure Storage Encryption. To avoid issues when installing SSL certificates, you should first synchronize the time between the following systems: • The server creating the certificates 86 | Physical Storage Management Guide
  • 87. • The key management servers • The storage system Requirements for SSL certificates Before obtaining and installing SSL certificates, you must understand what certificates are required and their requirements. SSL certificates for Storage Encryption must use the Privacy Enhanced Mail (PEM) Base-64 encoded X.509 format and follow a strict naming convention. The following table describes the required certificate types and naming conventions: Certificate for... Certificate type Certificate file name Storage system Public client.pem Storage system Private client_private.pem Key management server Public key_management_server_ipaddress_CA.pem key_management_server_ipaddress must be identical to the IP address of the key management server that you use to identify it when running the Storage Encryption setup program. These public and private certificates are required for the storage system and key management servers to establish secure SSL connections with each other and verify each other's identities. The certificates for the storage system are only used by the storage system's KMIP client. The private certificate can be passphrase protected during creation. In this case, the Storage Encryption setup program prompts you to enter the passphrase. If your key management server does not accept self-signed certificates, you also need to include the necessary certificate authority (CA) public certificate. In an HA pair, both nodes must use the same public and private certificates. If you want multiple HA pairs that are connected to the same key management server to have access to each other's keys, all nodes in all HA pairs must use the same public and private certificates. Installing SSL certificates on the storage system You install the necessary SSL certificates on the storage system by using the keymgr install cert command. The SSL certificates are required for securing the communication between the storage system and key management servers. Before you begin You must have obtained the public and private certificates for the storage system and the public certificate for the key management server and named them as required. Securing data at rest with Storage Encryption | 87
  • 88. Steps 1. Access the nodeshell: system node run -node node_name 2. Copy the certificate files to a temporary location on the storage system. 3. Install the public certificate of the storage system by entering the following command at the storage system prompt: keymgr install cert /path/client.pem 4. Install the private certificate of the storage system by entering the following command at the storage system prompt: keymgr install cert /path/client_private.pem 5. Install the public certificate of the key management server by entering the following command at the storage system prompt: keymgr install cert /path/key_management_server_ipaddress_CA.pem 6. If you are linking multiple key management servers to the storage system, repeat the previous step for each public certificate of each key management server. 7. Exit the nodeshell and return to the clustershell: exit Running the Storage Encryption setup wizard You launch the Storage Encryption setup wizard by using the key_manager setup command. You should run the Storage Encryption setup wizard after you complete setup of the storage system and the storage volumes or when you need to change Storage Encryption settings after initial setup. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. Enter the following command at the storage system prompt: key_manager setup 3. Complete the steps in the wizard to configure Storage Encryption. 4. Exit the nodeshell and return to the clustershell by entering the following command: exit 88 | Physical Storage Management Guide
  • 89. Example The following command launches the Storage Encryption setup wizard and shows an example of how to configure Storage Encryption: storage-system*> key_manager setup Found client certificate file client.pem. Registration successful for client.pem. Found client private key file client_private.pem. Is this file protected by a passphrase? [no]: Registration successful for client_private.pem. Enter the IP address for a key server, 'q' to quit: 172.22.192.192 Enter the IP address for a key server, 'q' to quit: q Enter the TCP port number for kmip server [6001] : You will now be prompted to enter a key tag name. The key tag name is used to identify all keys belonging to this Data ONTAP system. The default key tag name is based on the system's hostname. Would you like to use <storage-system> as the default key tag name? [yes]: Registering 1 key servers... Found client CA certificate file 172.22.192.192_CA.pem. Registration successful for 172.22.192.192_CA.pem. Registration complete. You will now be prompted for a subset of your network configuration setup. These parameters will define a pre-boot network environment allowing secure connections to the registered key server(s). Enter network interface: e0a Enter IP address: 172.16.132.165 Enter netmask: 255.255.252.0 Enter gateway: 172.16.132.1 Do you wish to enter or generate a passphrase for the system's encrypting drives at this time? [yes]: yes Would you like the system to autogenerate a passphrase? [yes]: yes Key ID: 080CDCB20000000001000000000000003FE505B0C5E3E76061EE48E02A29822C Make sure that you keep a copy of your passphrase, key ID, and key tag name in a secure location in case it is ever needed for recovery purposes. Should the system lock all encrypting drives at this time? yes Securing data at rest with Storage Encryption | 89
  • 90. Completed rekey on 4 disks: 4 successes, 0 failures, including 0 unknown key and 0 authentication failures. Completed lock on 4 disks: 4 successes, 0 failures, including 0 unknown key and 0 authentication failures. Managing Storage Encryption You can perform various tasks to manage Storage Encryption, including viewing and removing key management servers, and creating, deleting, restoring and synchronizing authentication keys. Adding key management servers You can use the key_manager add command to link key management servers to the storage system. This enables you to add additional key management servers for redundancy after initial setup or to replace existing key management servers. Before you begin You must first install the required storage system and key management server SSL certificates. If they are not present, the command fails. You must know the IP address for each key management server you want to link. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. To add a key management server, enter the following command: key_manager add -key_server key_server_ip_address 3. Exit the nodeshell and return to the clustershell by entering the following command: exit Example The following command adds a link from the storage system to the key management server with the IP address 172.16.132.118: storage-system> key_manager add -key_server 172.16.132.118 Found client certificate file client.pem. Registration successful for client.pem. Found client private key file client_private.pem. Is this file protected by a passphrase? [no]: no Registration successful for client_private.pem. 90 | Physical Storage Management Guide
  • 91. Registering 1 key servers... Found client CA certificate file 172.16.132.118_CA.pem. Registration successful for 172.16.132.118_CA.pem. Registration complete. Verifying key management server links You use the key_manager status or key_manager query commands to verify that all key management servers are successfully linked to the storage system. These commands are useful for verifying proper operation and troubleshooting. About this task Both commands display whether key management servers are responding. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. Perform one of the following actions: If you want to... Then enter the following command: Check the status of a specific key management server key_manager status - key_server key_server_ip_address Check the status of all key management servers key_manager status Check the status of all key management servers and view additional server details. key_manager query The key_manager query command displays additional information about key tags and key IDs. 3. Check the output to verify that all of the appropriate keys are available in the Data ONTAP key table. If the output of the key_manager query command displays key IDs marked with an asterisk (*), those keys exist on a key server but are not currently available in the Data ONTAP key table. To import those keys from the key management server into the key table, enter the following command: key_manager restore 4. Exit the nodeshell and return to the clustershell by entering the following command: exit Securing data at rest with Storage Encryption | 91
  • 92. Examples The following command checks the status of all key management servers linked to the storage system: storage-system> key_manager status Key server Status 172.16.132.118 Server is responding 172.16.132.211 Server is responding The following command checks the status of all key management servers linked to the storage system and displays additional information: storage-system> key_manager query Key server 172.16.132.118 is responding. Key server 172.16.132.211 is responding. Key server 172.16.132.118 reports 4 keys. Key tag Key ID -------- ------- storage-system 080CDCB20... storage-system 080CDCB20... storage-system 080CDCB20... storage-system 080CDCB20... Key server 172.16.132.211 reports 4 keys. Key tag Key ID -------- ------- storage-system *080CDCB20... storage-system 080CDCB20... storage-system 080CDCB20... storage-system *080CDCB20... Displaying key management server information You can display information about the external key management servers associated with the storage system by using the key_manager show command. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. To display external key management servers, enter the following command: key_manager show 92 | Physical Storage Management Guide
  • 93. All external key management servers associated with the storage system are listed. 3. Exit the nodeshell and return to the clustershell by entering the following command: exit Example The following command displays all external key management servers associated with the storage system: storage-system> key_manager show 172.18.99.175 Removing key management servers If you no longer want to use a key management server to store authentication keys used by self- encrypting disks in the storage system, you can remove the key management server link to the storage system by using the key_manager remove command. Before you begin You must know the IP address for each key management server that you want to remove. About this task Storage Encryption requires at least one key management server linked to the storage system to operate. If you want to replace a single key management server with another one, you must first add the new one before removing the old one. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. To remove key management servers, enter the following command: key_manager remove -key_server key_server_ip_address -key_server key_server_ip_address specifies the IP address of the key management server you want to remove. 3. Exit the nodeshell and return to the clustershell by entering the following command: exit Securing data at rest with Storage Encryption | 93
  • 94. Example The following command removes the link between the storage system and the key management server with the IP address 172.18.99.175: storage-system> key_manager remove -key_server 172.18.99.175 Key server 172.18.99.175 will be unregistered from service. Unregistration successful. What happens when key management servers are not reachable during the boot process Data ONTAP takes certain precautions to avoid undesired behavior in the event that the storage system cannot reach any of the specified key management servers during the boot process. If the storage system is configured for Storage Encryption, the SEDs are rekeyed and locked, and the SEDs are powered on, the storage system must retrieve the required authentication keys from the key management servers to authenticate itself to the SEDs before it can access the data. The storage system attempts to contact the specified key management servers for up to three hours. If the storage system cannot reach any of them after that time, the boot process stops and the storage system halts. If the storage system successfully contacts any specified key management server, it then attempts to establish an SSL connection for up to 15 minutes. If the storage system cannot establish an SSL connection with any specified key management server, the boot process stops and the storage system halts. While the storage system attempts to contact and connect to key management servers, it displays detailed information about the failed contact attempts at the CLI. You can interrupt the contact attempts at any time by pressing Ctrl-C. As a security measure, SEDs allow only a limited number of unauthorized access attempts, after which they disable access to the existing data. If the storage system cannot contact any specified key management servers to obtain the proper authentication keys, it can only attempt to authenticate with the default key which leads to a failed attempt and a panic. If the storage system is configured to automatically reboot in case of a panic, it enters a boot loop which results in continuous failed authentication attempts on the SEDs. Halting the storage system in these scenarios is by design to prevent the storage system from entering a boot loop and possible unintended data loss as a result of the SEDs locked permanently due to exceeding the safety limit of a certain number of consecutive failed authentication attempts. The limit and the type of lockout protection depends on the type of SED: 94 | Physical Storage Management Guide
  • 95. SED type Number of consecutive failed authentication attempts resulting in lockout Lockout protection type when safety limit is reached HDD 1024 Permanent. Data cannot be recovered, even when the proper authentication key becomes available again. SSD 5 Temporary. Lockout is only in effect until disk is power-cycled. For all SED types, a successful authentication resets the try count to zero. If you encounter this scenario where the storage system is halted due to failure to reach any specified key management servers, you must first identify and correct the cause for the communication failure before you attempt to continue booting the storage system. Displaying Storage Encryption disk information You can display information about self-encrypting disks by using the disk encrypt show command. This command displays the key ID and lock status for each self-encrypting disk. About this task The key ID displayed in the command output is an identifier used by Storage Encryption and key management servers as a reference to the authentication key. It is not the actual authentication key or the data encryption key. Steps 1. Access the nodeshell: system node run -node node_name 2. To display information about SEDs: disk encrypt show The disk encrypt show, lock, and rekey commands support extended wildcard matching. For more information, see the disk encrypt show man page. 3. Exit the nodeshell and return to the clustershell: exit Example The following command displays the status of each self-encrypting disk: Securing data at rest with Storage Encryption | 95
  • 96. storage-system> disk encrypt show Disk Key ID Locked? 0c.00.1 080CF0C8000000000100000000000000A948EE8604F4598ADFFB185B5BB7FED3 Yes 0c.00.0 080CF0C8000000000100000000000000A948EE8604F4598ADFFB185B5BB7FED3 Yes 0c.00.3 080CF0C8000000000100000000000000A948EE8604F4598ADFFB185B5BB7FED3 Yes 0c.00.4 080CF0C8000000000100000000000000A948EE8604F4598ADFFB185B5BB7FED3 Yes 0c.00.2 080CF0C8000000000100000000000000A948EE8604F4598ADFFB185B5BB7FED3 Yes 0c.00.5 080CF0C8000000000100000000000000A948EE8604F4598ADFFB185B5BB7FED3 Yes Changing the authentication key You can change the authentication key at any time by using the key_manager rekey command. You might want to change the authentication key as part of your security protocol or when moving an aggregate to another storage system. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. Perform one of the following actions: If you want to... Then... Change the authentication key and enter a new one manually a. Enter the following command at the storage system prompt: key_manager rekey -manual -key_tag key_tag b. When prompted, enter the new authentication key. It must be 20 to 32 characters long. Change the authentication key and have the system generate a new one automatically Enter the following command at the storage system prompt: key_manager rekey -key_tag key_tag key_tag is the label used to associate keys with a particular storage system. If you do not specify a key tag, the storage system uses the key tag specified when you set up Storage Encryption. If you did not specify this key tag during setup, it uses the parent key tag as the default. Each node has a parent key tag. HA pair members share the same parent key tag. 3. Exit the nodeshell and return to the clustershell by entering the following command: exit Example The following command changes the authentication key and prompts you to enter a new one manually. You can run the disk encrypt show command after completion to verify the results. 96 | Physical Storage Management Guide
  • 97. storage-system> key_manager rekey -manual Please enter a new passphrase: Please reenter the new passphrase: New passphrase generated. Key ID: 080CDCB2000000000100000000000000B0A11CBF3DDD20EFB0FBB5EE198DB22A Key tag: storage-system Notice: Remember to store the passphrase and the Key ID in a secure location. Passphrase, key ID, and key tag synchronized with the following key server(s): 172.16.132.118 172.16.132.211 Completed rekey on 4 disks: 4 successes, 0 failures, including 0 unknown key and 0 authentication failures. Retrieving authentication keys You can use the key_manager restore command to retrieve authentication keys from a key management server to a storage system. For example, when you created authentication keys on a node, you use this command to retrieve the keys for use on the partner node. Before you begin You must know the IP address for each key management server that you want to retrieve authentication keys from. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. To retrieve authentication keys from a key management server to the storage system, enter the following command: key_manager restore -key_server key_server_ip_address -key_tag key_tag If all specified key management servers are available, you can use the -all option instead of the -key_server option to clear out the current Data ONTAP key table and retrieve all keys matching the specified key tag from all specified key management servers. 3. Exit the nodeshell and return to the clustershell by entering the following command: exit Securing data at rest with Storage Encryption | 97
  • 98. Examples The following command restores keys with the key tag storage-system from the key management server with the IP address 172.18.99.175: storage-system> key_manager restore -key_server 172.18.99.175 - key_tag storage-system The following command restores all keys with the key tag storage-system from all key management servers linked to the storage system: storage-system> key_manager restore -all -key_tag storage-system Deleting an authentication key You can delete an authentication key that is no longer needed by removing it from the external key management server. Before you begin Verify that the authentication key is no longer needed before deleting it. Deleting an authentication key that is still in use can permanently prevent access to data on a storage system. Step 1. Refer to the documentation for the external key management server for details on how to delete stored authentication keys. SSL issues due to expired certificates If the SSL certificates used to secure key management communication between the storage system and key management servers expire, the storage system can no longer retrieve authentication keys from the key management server at bootup. This issue can cause data on SEDs to be unavailable. You can prevent this issue by updating all SSL certificates before their individual expiration dates. SSL certificates have a limited lifespan because they have an expiration date. After the SSL certificates reach their expiration dates, the certificates are no longer valid. When this happens, SSL connections that use expired certificates fail. For Storage Encryption, this means that the SSL connections between the storage system and the key management servers fail, the storage system no longer can retrieve authentication keys when needed, and data access to the SEDs fails, resulting in storage system panic and downtime. To prevent this issue from occurring, you must keep track of the expiration dates of all installed SSL certificates so that you can obtain new SSL certificates before they expire. 98 | Physical Storage Management Guide
  • 99. After you have obtained the new certificate files, you must first remove the existing certificate files from the storage system, and then install the new certificates on the storage system. Steps 1. Removing old SSL certificates before installing new ones on page 99 If you want to update or reinstall the SSL certificates used by Storage Encryption, you must first manually remove the old ones to ensure that the new ones are used. 2. Installing replacement SSL certificates on the storage system on page 99 After you remove the old certificates, you create the new replacement SSL certificates, save them with the proper file name and format, and then install them on the storage system. Removing old SSL certificates before installing new ones If you want to update or reinstall the SSL certificates used by Storage Encryption, you must first manually remove the old ones to ensure that the new ones are used. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. Remove the IP addresses of all key management servers by entering the following command for each key management server: key_manager remove -key_server key_server_ip_address 3. Remove the storage system's client certificates by entering the following commands: keymgr delete cert client_private.pem keymgr delete cert client.pem 4. Remove all installed key management server certificates by entering the following commands for each key management server: keymgr delete cert key_server_ip_address_CA.pem 5. Exit the nodeshell and return to the clustershell by entering the following command: exit Installing replacement SSL certificates on the storage system After you remove the old certificates, you create the new replacement SSL certificates, save them with the proper file name and format, and then install them on the storage system. Before you begin • You must have removed the old certificates that are about to expire from the storage system. Securing data at rest with Storage Encryption | 99
  • 100. • You must have obtained the replacement public and private certificates for the storage system and the public certificate for the key management server, and named them as required. For more information, see the Clustered Data ONTAP System Administration Guide for Cluster Administrators. • You must have installed the appropriate new certificates on the key management server. For more information, see the documentation for your key management server. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. Copy the certificate files to a temporary location on the storage system. 3. Install the public certificate of the storage system by entering the following command: keymgr install cert /path/client.pem 4. Install the private certificate of the storage system by entering the following command: keymgr install cert /path/client_private.pem 5. Install the public certificate of all key management servers by entering the following command for each key management server: keymgr install cert /path/key_management_server_ipaddress_CA.pem 6. Add all key management servers by entering the following command for each key management server: key_manager add -key_server key_server_ip_address 7. Verify connectivity between the storage system and key management servers by entering the following command: key_manager query You should see a list of existing key IDs retrieved from the key management servers. 8. Exit the nodeshell and return to the clustershell by entering the following command: exit Returning SEDs to unprotected mode If your storage system is configured to use Storage Encryption but you decide to stop using this feature, you can do so by returning the SEDs to unprotected mode. You cannot disable Storage Encryption altogether because SEDs always encrypt data for storage. However, you can return them 100 | Physical Storage Management Guide
  • 101. to unprotected mode where they no longer use secret authentication keys, and use the default MSID instead. Steps 1. Access the nodeshell by entering the following command: system node run -node node_name 2. To change the authentication key for all SEDs on the storage system back to the default MSID, enter the following command: disk encrypt rekey * 0x0 3. If you expect to operate the storage system in unprotected mode permanently, you should also remove all key management servers by entering the following command for each one: key_manager remove -key_server key_server_ip_address -key_server key_server_ip_address specifies the IP address of the key management server you want to remove. The storage system displays two kmip_init errors during every bootup after you remove all key management servers. These errors are normal in this situation and you can disregard them. 4. If you expect to operate the storage system in unprotected mode permanently and you removed all key management servers in the preceding step, you should view the list of installed Storage Encryption related SSL certificates, and then remove all key management server SSL certificates: keymgr cert list keymgr delete cert client.pem keymgr delete cert client_private.pem keymgr delete cert key_management_server_ipaddress_CA.pem If you had multiple key management servers linked to the storage system, repeat the last command for each public certificate of each key management server. 5. Exit the nodeshell and return to the clustershell by entering the following command: exit Securing data at rest with Storage Encryption | 101
  • 102. Destroying data on disks using Storage Encryption You can destroy data stored on disks using Storage Encryption for security reasons, including sanitizing the disks, setting the disk state to end-of-life, and emergency shredding of the data. Sanitizing disks using Storage Encryption before return to vendor If you want to return a disk to a vendor but do not want anyone to access sensitive data on the disk, you can sanitize it first by using the disk encrypt sanitize command. This renders the data on the disk inaccessible, but the disk can be reused. This command only works on spare disks. Steps 1. Migrate any data that needs to be preserved to a different aggregate. 2. Destroy the aggregate. 3. Access the nodeshell by entering the following command: system node run -node node_name 4. Identify the disk ID for the disk to be sanitized by entering the following command: disk encrypt show 5. Enter the following command: disk encrypt sanitize disk_ID 6. Exit the nodeshell and return to the clustershell by entering the following command: exit Example The following command sanitizes a self-encrypting disk with the disk ID 0c.00.3. You can run the sysconfig -r command before and after the operation to verify the results. storage-system> sysconfig -r Aggregate aggr0 (online, raid_dp) (block checksums) Plex /aggr0/plex0 (online, normal, active) RAID group /aggr0/plex0/rg0 (normal) RAID Disk Device HA SHELF BAY CHAN Pool Type RPM Used (MB/blks) Phys (MB/blks) --------- ------ ------------- ---- ---- ---- ----- -------------- -------------- dparity 0c.00.0 0c 0 0 SA:B - SAS 15000 560000/1146880000 560208/1147307688 parity 0c.00.1 0c 0 1 SA:B - SAS 15000 560000/1146880000 560208/1147307688 data 0c.00.2 0c 0 2 SA:B - SAS 15000 560000/1146880000 560208/1147307688 Spare disks RAID Disk Device HA SHELF BAY CHAN Pool Type RPM Used (MB/blks) Phys (MB/blks) --------- ------ ------------- ---- ---- ---- ----- -------------- -------------- Spare disks for block or zoned checksum traditional volumes or aggregates spare 0c.00.3 0c 0 3 SA:B - SAS 15000 560000/1146880000 560208/1147307688 spare 0c.00.4 0c 0 4 SA:B - SAS 15000 560000/1146880000 560208/1147307688 102 | Physical Storage Management Guide
  • 103. spare 0c.00.5 0c 0 5 SA:B - SAS 15000 560000/1146880000 560208/1147307688 storage-system> disk encrypt sanitize 0c.00.3 storage-system> Wed Jun 30 17:49:16 PDT [disk.failmsg:error]: Disk 0c.00.3 (3SL04F3V00009015WTHU): message received. Wed Jun 30 17:49:16 PDT [raid.disk.unload.done:info]: Unload of Disk 0c.00.3 Shelf 0 Bay 3 [SYSTEM X415_S15K7560A15 NQS3] S/N [3SL04F3V00009015WTHU] has completed successfully storage-system> Wed Jun 30 17:49:25 PDT [disk.sanit.complete:info]: Disk 0c.00.3 [S/N 3SL04F3V00009015WTHU] has completed sanitization. storage-system> sysconfig - r Aggregate aggr0 (online, raid_dp) (block checksums) Plex /aggr0/plex0 (online, normal, active) RAID group /aggr0/plex0/rg0 (normal) RAID Disk Device HA SHELF BAY CHAN Pool Type RPM Used (MB/blks) Phys (MB/blks) --------- ------ ------------- ---- ---- ---- ----- -------------- -------------- dparity 0c.00.0 0c 0 0 SA:B - SAS 15000 560000/1146880000 560208/1147307688 parity 0c.00.1 0c 0 1 SA:B - SAS 15000 560000/1146880000 560208/1147307688 data 0c.00.2 0c 0 2 SA:B - SAS 15000 560000/1146880000 560208/1147307688 Spare disks RAID Disk Device HA SHELF BAY CHAN Pool Type RPM Used (MB/blks) Phys (MB/blks) --------- ------ ------------- ---- ---- ---- ----- -------------- -------------- Spare disks for block or zoned checksum traditional volumes or aggregates spare 0c.00.4 0c 0 4 SA:B - SAS 15000 560000/1146880000 560208/1147307688 spare 0c.00.5 0c 0 5 SA:B - SAS 15000 560000/1146880000 560208/1147307688 Maintenance disks RAID Disk Device HA SHELF BAY CHAN Pool Type RPM Used (MB/blks) Phys (MB/blks) --------- ------ ------------- ---- ---- ---- ----- -------------- -------------- sanitized 0c.00.3 0c 0 3 SA:B - SAS 15000 560000/1146880000 560208/1147307688 storage-system> Setting the state of disks using Storage Encryption to end-of-life If you want to render a disk permanently unusable and the data on it inaccessible, you can set the state of the disk to end-of-life by using the disk encrypt destroy command. This command only works on spare disks. Steps 1. Remove any data from the aggregate containing the disk. 2. Migrate any data that needs to be preserved to a different aggregate. 3. Destroy the aggregate. 4. Access the nodeshell: system node run -node node_name 5. Enter the following command: disk encrypt destroy disk_ID 6. Exit the nodeshell and return to the clustershell: exit Securing data at rest with Storage Encryption | 103
  • 104. 7. If the disk has a PSID printed on its label but you do not want the disk to be able to be reset to factory settings and returned to service at a later time, obliterate all instances of the PSID on the disk label (text or scannable code). Be sure to also destroy any copies, scans, or photographs of the disk label. Result The disk's encryption key is set to an unknown random value and the disk is irreversibly locked. The disk is now completely unusable and can be safely disposed of without risk of unauthorized data access. Emergency shredding of data on disks using Storage Encryption In case of a security emergency, you can instantly prevent access to data on disks using Storage Encryption, even if power is not available to the storage system or the external key server. Before you begin You must configure the external key server so that it only operates if an easily destroyed authentication item (for example, a smart card or USB drive) is present. See the documentation for the external key management server for details. About this task The steps for emergency shredding vary depending on whether power is available to the storage system and the external key server. Step 1. Perform one of the following actions: If... Then... Power is available to the storage system and you have time to gracefully take the storage system offline a. If the storage system is a node in an HA pair, disable takeover. b. Take all aggregates offline and destroy them. c. Halt the storage system. d. Boot into maintenance mode. e. Enter the following command: disk encrypt sanitize -all This leaves the storage system in a permanently disabled state with all data erased. To use the storage system again, you must set it up from the beginning. 104 | Physical Storage Management Guide
  • 105. If... Then... Power is available to the storage system and you must shred the data immediately; time is critical a. If the storage system is a node in an HA pair, disable takeover. b. Access the nodeshell by entering the following command: system node run -node node_name c. Set the privilege level to advanced. d. Enter the following command: disk encrypt sanitize -all The storage system panics, which is expected due to the abrupt nature of the procedure. It leaves the storage system in a permanently disabled state with all data erased. To use the storage system again, you must set it up from the beginning. Power is available to the external key server but not to the storage system a. Log in to the external key server. b. Destroy all keys associated with the disks containing data to protect. Power is not available to the external key server or the storage system Destroy the authentication item for the key server (for example, the smart card). If power to the systems is restored, the external key server cannot operate due to the missing authentication item. This prevents access to the disk encryption keys by the storage system, and therefore access to the data on the disks. What function the physical secure ID has for SEDs Certain self-encrypting disks (SEDs) feature additional functionality to reset the disk to factory settings. These disks have a physical secure ID (PSID) printed on the disk label that is required to perform a factory reset. The PSID is unique to each drive. It is printed on the disk label and visible to anyone with physical access to the SED. The PSID is not electronically retrievable from the SED. If the disk label is obliterated, the PSID is lost and cannot be recovered. Using the PSID to perform a factory reset causes all disk parameters to be reset to factory original settings, including the following: • The encryption key used to encrypt and decrypt the data on the media is changed to an unknown value. If the SED contained data, access to the data is permanently lost. The new unknown encryption key cannot be retrieved. This operation cannot be undone. • The authentication key required to authenticate to the SED is changed back to the manufacturer's default secure ID (MSID). New data stored on the SED is not protected until the MSID is changed to a new secret authentication key. Securing data at rest with Storage Encryption | 105
  • 106. • The SED can be returned into service, even if its state was previously set to end-of-life. If the SED was previously used but then set to end-of-life state using the disk encrypt destroy command, use of the PSID recovers the SED from this state and returns it to normal service state. However, it only returns it to factory original settings. It cannot in any way recover previously used encryption or authentication keys or restore access to previous data stored on the SED. SEDs that have PSID functionality There are several models of SEDs but only some of them have PSID functionality. Earlier SED models with PSID functionality do not have the PSID printed on the disk label and therefore cannot use it. Use the following table for guidance to find out whether your SEDs have PSID functionality: Disk model Has PSID functio nality Has PSID printed on disk label Comments X414 No No SED cannot be reset to factory original settings using a PSID. X415 No No SED cannot be reset to factory original settings using a PSID. X416 Yes Maybe Check the label on the physical disk. If a PSID is printed on the label, the SED can be reset to factory original settings using the PSID. X417 Yes Maybe Check the label on the physical disk. If a PSID is printed on the label, the SED can be reset to factory original settings using the PSID. All other SEDs Yes Yes Check the label on the physical disk to obtain the PSID. Resetting an SED to factory original settings If you previously set the state of an SED to end-of-life by using the disk encrypt destroy command but now want to return it to service, you can reset it to its factory original settings by using the disk encrypt revert_original command provided that the disk has a PSID printed on its label. Before you begin You must have obtained the SED's PSID from its disk label. 106 | Physical Storage Management Guide
  • 107. Steps 1. Access the nodeshell: system node run -node node_name 2. Set the privilege level to advanced: priv set advanced 3. Reset the disk to its factory configured settings: disk encrypt revert_original disk_ID PSID The PSID is the physical secure ID printed on the disk label that is required to perform a factory reset. 4. Make the disk unavailable to the storage system: disk fail disk_name 5. Make the disk available again to the storage system: disk unfail -s disk_name 6. Verify that the operation was successful: sysconfig -r The disk that you reset shows up as a spare disk. 7. Return to the admin privilege level: priv set admin 8. Exit the nodeshell and return to the clustershell: exit Securing data at rest with Storage Encryption | 107
  • 108. How Data ONTAP uses RAID to protect your data and data availability Understanding how RAID protects your data and data availability can help you administer your storage systems more effectively. For native storage, Data ONTAP uses RAID-DP (double-parity) or RAID Level 4 (RAID4) protection to ensure data integrity within a RAID group even if one or two of those drives fail. Parity drives provide redundancy for the data stored in the data drives. If a drive fails (or, for RAID-DP, up to two drives), the RAID subsystem can use the parity drives to reconstruct the data in the drive that failed. For array LUNs, Data ONTAP stripes data across the array LUNs using RAID0. The storage arrays, not Data ONTAP, provide the RAID protection for the array LUNs that they make available to Data ONTAP. RAID protection levels for disks Data ONTAP supports two levels of RAID protection for aggregates composed of disks in native disk shelves: RAID-DP and RAID4. RAID-DP is the default RAID level for new aggregates. For more information about configuring RAID, see Technical Report 3437: Storage Subsystem Resiliency Guide. Related information TR 3437: Storage Subsystem Resiliency Guide What RAID-DP protection is If an aggregate is configured for RAID-DP protection, Data ONTAP reconstructs the data from one or two failed disks within a RAID group and transfers that reconstructed data to one or two spare disks as necessary. RAID-DP provides double-parity disk protection when the following conditions occur: • There is a single-disk failure or double-disk failure within a RAID group. • There are media errors on a block when Data ONTAP is attempting to reconstruct a failed disk. The minimum number of disks in a RAID-DP group is three: at least one data disk, one regular parity disk, and one double-parity (dParity) disk. However, for non-root aggregates with only one RAID group, you must have at least 5 disks (three data disks and two parity disks). If there is a data-disk failure or parity-disk failure in a RAID-DP group, Data ONTAP replaces the failed disk in the RAID group with a spare disk and uses the parity data to reconstruct the data of the 108 | Physical Storage Management Guide
  • 109. failed disk on the replacement disk. If there is a double-disk failure, Data ONTAP replaces the failed disks in the RAID group with two spare disks and uses the double-parity data to reconstruct the data of the failed disks on the replacement disks. RAID-DP is the default RAID type for all aggregates. What RAID4 protection is RAID4 provides single-parity disk protection against single-disk failure within a RAID group. If an aggregate is configured for RAID4 protection, Data ONTAP reconstructs the data from a single failed disk within a RAID group and transfers that reconstructed data to a spare disk. The minimum number of disks in a RAID4 group is two: at least one data disk and one parity disk. However, for non-root aggregates with only one RAID group, you must have at least 3 disks (two data disks and one parity disk). If there is a single data or parity disk failure in a RAID4 group, Data ONTAP replaces the failed disk in the RAID group with a spare disk and uses the parity data to reconstruct the failed disk’s data on the replacement disk. If no spare disks are available, Data ONTAP goes into degraded mode and alerts you of this condition. Attention: With RAID4, if there is a second disk failure before data can be reconstructed from the data on the first failed disk, there will be data loss. To avoid data loss when two disks fail, you can select RAID-DP. This provides two parity disks to protect you from data loss when two disk failures occur in the same RAID group before the first failed disk can be reconstructed. RAID protection for array LUNs Storage arrays provide the RAID protection for the array LUNs that they make available to Data ONTAP; Data ONTAP does not provide the RAID protection. Data ONTAP uses RAID0 (striping) for array LUNs. Data ONTAP supports a variety of RAID types on the storage arrays, except RAID0 because RAID0 does not provide storage protection. When creating RAID groups on storage arrays, you need to follow the best practices of the storage array vendor to ensure that there is an adequate level of protection on the storage array so that disk failure does not result in loss of data or loss of access to data. Note: A RAID group on a storage array is the arrangement of disks that together form the defined RAID level. Each RAID group supports only one RAID type. The number of disks that you select for a RAID group determines the RAID type that a particular RAID group supports. Different storage array vendors use different terms to describe this entity—RAID groups, parity groups, disk groups, Parity RAID groups, and other terms. Note: Data ONTAP supports RAID4 and RAID-DP on native disk shelves but supports only RAID0 on array LUNs. How Data ONTAP uses RAID to protect your data and data availability | 109
  • 110. RAID protection for Data ONTAP-v storage Because Data ONTAP-v storage is connected to the host server, rather than a storage system running Data ONTAP, the host server provides the RAID protection for the physical disks. Data ONTAP uses RAID0 for the virtual disks to optimize performance. See the Data ONTAP Edge Installation and Administration Guide for more information. Protection provided by RAID and SyncMirror Combining RAID and SyncMirror provides protection against more types of drive failures than using RAID alone. You can use RAID in combination with the SyncMirror functionality, which also offers protection against data loss due to drive or other hardware component failure. SyncMirror protects against data loss by maintaining two copies of the data contained in the aggregate, one in each plex. Any data loss due to drive failure in one plex is repaired by the undamaged data in the other plex. For more information about SyncMirror, see the Clustered Data ONTAP Data Protection Guide. The following tables show the differences between using RAID alone and using RAID with SyncMirror: Table 1: RAID-DP and SyncMirror Criteria RAID-DP alone RAID-DP with SyncMirror Failures protected against • Single-drive failure • Double-drive failure within a single RAID group • Multiple-drive failures, as long as no more than two drives within a single RAID group fail • All failures protected against by RAID-DP alone • Any combination of failures protected against by RAID-DP alone in one plex, concurrent with an unlimited number of failures in the other plex • Storage subsystem failures (HBA, cables, shelf), as long as only one plex is affected 110 | Physical Storage Management Guide
  • 111. Criteria RAID-DP alone RAID-DP with SyncMirror Failures not protected against • Three or more concurrent drive failures within a single RAID group • Storage subsystem failures (HBA, cables, shelf) that lead to three or more concurrent drive failures within a single RAID group • Three or more concurrent drive failures in a single RAID group on both plexes Required resources per RAID group n data drives + 2 parity disks 2 x (n data drives + 2 parity drives) Performance cost Almost none Low mirroring overhead; can improve performance Additional cost and complexity None SyncMirror license and configuration Table 2: RAID4 and SyncMirror Criteria RAID4 alone RAID4 with SyncMirror Failures protected against • Single-disk failure • Multiple-disk failures, as long as no more than one disk within a single RAID group fails • All failures protected against by RAID4 alone • Any combination of failures protected against by RAID4 alone in one plex, concurrent with an unlimited number of failures in the other plex • Storage subsystem failures (HBA, cables, shelf), as long as only one plex is affected How Data ONTAP uses RAID to protect your data and data availability | 111
  • 112. Criteria RAID4 alone RAID4 with SyncMirror Failures not protected against • Two or more concurrent drive failures within a single RAID group • Storage subsystem failures (HBA, cables, shelf) that lead to two or more concurrent drive failures within a single RAID group • Two or more concurrent drive failures in a single RAID group on both plexes Required resources per RAID group n data drives + 1 parity drive 2 x (n data drives + 1 parity drive) Performance cost None Low mirroring overhead; can improve performance Additional cost and complexity None SyncMirror configuration and extra storage requirement Table 3: RAID0 and SyncMirror Criteria RAID0 alone RAID0 with SyncMirror Failures protected against No protection against any failures RAID protection is provided by the RAID implemented on the storage array. Any combination of array LUN, connectivity, or hardware failures, as long as only one plex is affected Failures not protected against No protection against any failures RAID protection is provided by the RAID implemented on the storage array. Any concurrent failures that affect both plexes Required array LUN resources per RAID group No extra array LUNs required other than n data array LUNs 2 x n data array LUNs Performance cost None Low mirroring overhead; can improve performance Additional cost and complexity None SyncMirror configuration and extra storage requirement 112 | Physical Storage Management Guide
  • 113. Understanding RAID drive types Data ONTAP classifies drives (or, for partitioned drives, partitions) as one of four types for RAID: data, hot spare, parity, or dParity. You manage disks differently depending on whether they are spare or being used in an aggregate. The RAID type is determined by how RAID is using a drive or partition; it is different from the Data ONTAP disk type. You cannot affect the RAID type for a drive. The RAID type is displayed in the Position column for many storage commands. For drives using root-data partitioning and SSDs in storage pools, a single drive might be used in multiple ways for RAID. For example, the root partition of a partitioned drive might be a spare partition, whereas the data partition might be being used for parity. For this reason, the RAID drive type for partitioned drives and SSDs in storage pools is displayed simply as shared. Data disk Holds data stored on behalf of clients within RAID groups (and any data generated about the state of the storage system as a result of a malfunction). Spare disk Does not hold usable data, but is available to be added to a RAID group in an aggregate. Any functioning disk that is not assigned to an aggregate but is assigned to a system functions as a hot spare disk. Parity disk Stores row parity information that is used for data reconstruction when a single disk drive fails within the RAID group. dParity disk Stores diagonal parity information that is used for data reconstruction when two disk drives fail within the RAID group, if RAID-DP is enabled. How RAID groups work A RAID group consists of one or more data disks or array LUNs, across which client data is striped and stored, and up to two parity disks, depending on the RAID level of the aggregate that contains the RAID group. RAID-DP uses two parity disks to ensure data recoverability even if two disks within the RAID group fail. RAID4 uses one parity disk to ensure data recoverability if one disk within the RAID group fails. How Data ONTAP uses RAID to protect your data and data availability | 113
  • 114. RAID0 does not use any parity disks; it does not provide data recoverability if any disks within the RAID group fail. How RAID groups are named Within each aggregate, RAID groups are named rg0, rg1, rg2, and so on in order of their creation. You cannot specify the names of RAID groups. Considerations for sizing RAID groups Configuring an optimum RAID group size requires a trade-off of factors. You must decide which factors—speed of RAID rebuild, assurance against risk of data loss due to drive failure, optimizing I/O performance, and maximizing data storage space—are most important for the aggregate that you are configuring. When you create larger RAID groups, you maximize the space available for data storage for the same amount of storage used for parity (also known as the “parity tax”). On the other hand, when a disk fails in a larger RAID group, reconstruction time is increased, impacting performance for a longer period of time. In addition, having more disks in a RAID group increases the probability of a multiple disk failure within the same RAID group. HDD or array LUN RAID groups You should follow these guidelines when sizing your RAID groups composed of HDDs or array LUNs: • All RAID groups in an aggregate should have a similar number of disks. The RAID groups do not have to be exactly the same size, but you should avoid having any RAID group that is less than one half the size of other RAID groups in the same aggregate when possible. • The recommended range of RAID group size is between 12 and 20. The reliability of performance disks can support a RAID group size of up to 28, if needed. • If you can satisfy the first two guidelines with multiple RAID group sizes, you should choose the larger size. SSD RAID groups in Flash Pool aggregates The SSD RAID group size can be different from the RAID group size for the HDD RAID groups in a Flash Pool aggregate. Usually, you should ensure that you have only one SSD RAID group for a Flash Pool aggregate, to minimize the number of SSDs required for parity. SSD RAID groups in SSD aggregates You should follow these guidelines when sizing your RAID groups composed of SSDs: • All RAID groups in an aggregate should have a similar number of drives. 114 | Physical Storage Management Guide
  • 115. The RAID groups do not have to be exactly the same size, but you should avoid having any RAID group that is less than one half the size of other RAID groups in the same aggregate when possible. • For RAID-DP, the recommended range of RAID group size is between 20 and 28. Related references Storage limits on page 195 Customizing the size of your RAID groups You can customize the size of your RAID groups to ensure that your RAID group sizes are appropriate for the amount of storage you plan to include for an aggregate. About this task For standard aggregates, you change the size of RAID groups on a per-aggregate basis. For Flash Pool aggregates, you can change the RAID group size for the SSD RAID groups and the HDD RAID groups independently. The following list outlines some facts about changing the RAID group size: • By default, if the number of disks or array LUNs in the most recently created RAID group is less than the new RAID group size, disks or array LUNs will be added to the most recently created RAID group until it reaches the new size. • All other existing RAID groups in that aggregate remain the same size, unless you explicitly add disks to them. • You can never cause a RAID group to become larger than the current maximum RAID group size for the aggregate. • You cannot decrease the size of already created RAID groups. • The new size applies to all RAID groups in that aggregate (or, in the case of a Flash Pool aggregate, all RAID groups for the affected RAID group type—SSD or HDD). Step 1. Use the applicable command: If you want to... Enter the following command... Change the maximum RAID group size for the SSD RAID groups of a Flash Pool aggregate storage aggregate modify -aggregate aggr_name - cache-raid-group-size size How Data ONTAP uses RAID to protect your data and data availability | 115
  • 116. If you want to... Enter the following command... Change the maximum size of any other RAID groups storage aggregate modify -aggregate aggr_name - maxraidsize size Examples The following command changes the maximum RAID group size of the aggregate n1_a4 to 20 disks or array LUNs: storage aggregate modify -aggregate n1_a4 -maxraidsize 20 The following command changes the maximum RAID group size of the SSD cache RAID groups of the Flash Pool aggregate n1_cache_a2 to 24: storage aggregate modify -aggregate n1_cache_a2 -cache-raid-group-size 24 Related concepts Considerations for sizing RAID groups on page 114 Considerations for Data ONTAP RAID groups for array LUNs Setting up Data ONTAP RAID groups for array LUNs requires planning and coordination with the storage array administrator so that the administrator makes the number and size of array LUNs you need available to Data ONTAP. For array LUNs, Data ONTAP uses RAID0 RAID groups to determine where to allocate data to the LUNs on the storage array. The RAID0 RAID groups are not used for RAID data protection. The storage arrays provide the RAID data protection. Note: Data ONTAP RAID groups are similar in concept to what storage array vendors call RAID groups, parity groups, disk groups, Parity RAID groups, and other terms. Follow these steps when planning your Data ONTAP RAID groups for array LUNs: 1. Plan the size of the aggregate that best meets your data needs. 2. Plan the number and size of RAID groups that you need for the size of the aggregate. Note: It is best to use the default RAID group size for array LUNs. The default RAID group size is adequate for most organizations. The default RAID group size is different for array LUNs and disks. 3. Plan the size of the LUNs that you need in your RAID groups. • To avoid a performance penalty, all array LUNs in a particular RAID group should be the same size. 116 | Physical Storage Management Guide
  • 117. • The LUNs should be the same size in all RAID groups in the aggregate. 4. Ask the storage array administrator to create the number of LUNs of the size you need for the aggregate. The LUNs should be optimized for performance, according to the instructions in the storage array vendor documentation. 5. Create all the RAID groups in the aggregate at the same time. Note: Do not mix array LUNs from storage arrays with different characteristics in the same Data ONTAP RAID group. Note: If you create a new RAID group for an existing aggregate, be sure that the new RAID group is the same size as the other RAID groups in the aggregate, and that the array LUNs are the same size as the LUNs in the other RAID groups in the aggregate. How Data ONTAP works with hot spare disks A hot spare disk is a disk that is assigned to a storage system but is not in use by a RAID group. It does not yet hold data but is ready for use. If a disk failure occurs within a RAID group, Data ONTAP automatically assigns hot spare disks to RAID groups to replace the failed disks. Minimum number of hot spares you should have Having insufficient spares increases the risk of a disk failure with no available spare, resulting in a degraded RAID group. A spare disk is also required to provide important information (a core file) to technical support in case of a controller disruption. MSATA disks, or disks in a multi-disk carrier, should have four hot spares during steady state operation, and you should never allow the number of MSATA hot spares to dip below two. For RAID groups composed of SSDs, you should have at least one spare disk. For all other Data ONTAP disk types, you should have at least one matching or appropriate hot spare available for each kind of disk installed in your storage system. However, having two available hot spares for all disks provides the best protection against disk failure. Having at least two available hot spares provides the following benefits: • When you have two or more hot spares for a data disk, Data ONTAP can put that disk into the maintenance center if needed. Data ONTAP uses the maintenance center to test suspect disks and take offline any disk that shows problems. • Having two hot spares means that when a disk fails, you still have a spare available if another disk fails before you replace the first failed disk. How Data ONTAP uses RAID to protect your data and data availability | 117
  • 118. A single spare disk can serve as a hot spare for multiple RAID groups. However, if any disk in those RAID groups fails, then no spare is available for any future disk failures, or for a core file, until the spare is replaced. For this reason, having more than one spare is recommended. Related concepts Spare requirements for multi-disk carrier disks on page 28 What disks can be used as hot spares A disk must conform to certain criteria to be used as a hot spare for a particular data disk. For a disk to be used as a hot spare for another disk, it must conform to the following criteria: • It must be either an exact match for the disk it is replacing or an appropriate alternative. • If SyncMirror is in use, the spare must be in the same pool as the disk it is replacing. • The spare must be owned by the same system as the disk it is replacing. What a matching spare is A matching hot spare exactly matches several characteristics of a designated data disk. Understanding what a matching spare is, and how Data ONTAP selects spares, enables you to optimize your spares allocation for your environment. A matching spare is a disk that exactly matches a data disk for all of the following criteria: • Effective Data ONTAP disk type The effective disk type can be affected by the value of the raid.mix.hdd.performance and raid.mix.hdd.capacity options, which determine the disk types that are considered to be equivalent. • Size • Speed (RPM) • Checksum type (BCS or AZCS) Related concepts How Data ONTAP reports disk types on page 10 What an appropriate hot spare is If a disk fails and no hot spare disk that exactly matches the failed disk is available, Data ONTAP uses the best available spare. Understanding how Data ONTAP chooses an appropriate spare when there is no matching spare enables you to optimize your spare allocation for your environment. Data ONTAP picks a non-matching hot spare based on the following criteria: 118 | Physical Storage Management Guide
  • 119. • If the available hot spares are not the correct size, Data ONTAP uses one that is the next size up, if there is one. The replacement disk is downsized to match the size of the disk it is replacing; the extra capacity is not available. • If the available hot spares are not the correct speed, Data ONTAP uses one that is a different speed. Using drives with different speeds within the same aggregate is not optimal. Replacing a disk with a slower disk can cause performance degradation, and replacing a disk with a faster disk is not cost-effective. • If the failed disk is part of a mirrored aggregate and there are no hot spares available in the correct pool, Data ONTAP uses a spare from the other pool. Using drives from the wrong pool is not optimal because you no longer have fault isolation for your SyncMirror configuration. If no spare exists with an equivalent disk type or checksum type, the RAID group that contains the failed disk goes into degraded mode; Data ONTAP does not combine effective disk types or checksum types within a RAID group. Related concepts How Data ONTAP reports disk types on page 10 About degraded mode When a disk fails, Data ONTAP can continue to serve data, but it must reconstruct the data from the failed disk using RAID parity. When this happens, the affected RAID group is said to be in degraded mode. The performance of a storage system with one or more RAID groups in degraded mode is decreased. A RAID group goes into degraded mode in the following scenarios: • A single disk fails in a RAID4 group. After the failed disk is reconstructed to a spare, the RAID group returns to normal mode. • One or two disks fail in a RAID-DP group. If two disks have failed in a RAID-DP group, the RAID group goes into double-degraded mode. • A disk is taken offline by Data ONTAP. After the offline disk is brought back online, the RAID group returns to normal mode. Note: If another disk fails in a RAID-DP group in double-degraded mode or a RAID4 group in degraded mode, data loss could occur (unless the data is mirrored). For this reason, always minimize the amount of time a RAID group is in degraded mode by ensuring that appropriate hot spares are available. How Data ONTAP uses RAID to protect your data and data availability | 119
  • 120. Related concepts How Data ONTAP handles a failed disk that has no available hot spare on page 121 How low spare warnings can help you manage your spare drives By default, Data ONTAP issues warnings to the console and logs if you have fewer than one hot spare drive that matches the attributes of each drive in your storage system. You can change the threshold value for these warning messages to ensure that your system adheres to best practices. To make sure that you always have two hot spares for every drive (a best practice), you can set the min_spare_count RAID option to 2. Setting the min_spare_count RAID option to 0 disables low spare warnings. You might want to do this if you do not have enough drives to provide hot spares (for example, if your storage system does not support external disk shelves). You can disable the warnings only if the following requirements are met: • Your system has 16 or fewer drives. • You have no RAID groups that use RAID4. Note: You cannot create aggregates that use RAID4 protection while the raid.min_spare_count option is set to 0. If either of these requirements is no longer met after this option has been set to 0, the option is automatically set back to 1. How Data ONTAP handles a failed disk with a hot spare Using an available matching hot spare, Data ONTAP can use RAID to reconstruct the missing data from the failed disk onto the hot spare disk with no data service interruption. If a disk fails and a matching or appropriate spare is available, Data ONTAP performs the following tasks: • Replaces the failed disk with a hot spare disk. If RAID-DP is enabled and a double-disk failure occurs in the RAID group, Data ONTAP replaces each failed disk with a separate spare disk. • In the background, reconstructs the missing data onto the hot spare disk or disks. Note: During reconstruction, the system is in degraded mode, and file service might slow down. • Logs the activity in the /etc/messages file. • Sends an AutoSupport message. Attention: Always replace the failed disks with new hot spare disks as soon as possible, so that hot spare disks are always available in the storage system. 120 | Physical Storage Management Guide
  • 121. Note: If the available spare disks are not the correct size, Data ONTAP chooses a disk of the next larger size and restricts its capacity to match the size of the disk it is replacing. Related concepts How Data ONTAP handles a failed disk that has no available hot spare on page 121 How Data ONTAP handles a failed disk that has no available hot spare When a failed disk has no appropriate hot spare available, Data ONTAP puts the affected RAID group into degraded mode indefinitely and the storage system automatically shuts down within a specified time period. If the maximum number of disks have failed in a RAID group (two for RAID-DP, one for RAID4), the storage system automatically shuts down in the period of time specified by the raid.timeout option. The default timeout value is 24 hours. To ensure that you are aware of the situation, Data ONTAP sends an AutoSupport message whenever a disk fails. In addition, it logs a warning message in the /etc/message file once per hour after a disk fails. Attention: If a disk fails and no hot spare disk is available, contact technical support. Related concepts How Data ONTAP handles a failed disk with a hot spare on page 120 About degraded mode on page 119 Considerations for changing the timeout RAID option The raid.timeout option controls how long a storage system runs after a RAID group goes into degraded mode or the NVRAM battery malfunctions or loses power. You can change the value of this option, but you should understand the implications of doing so. The purpose for the system shutdown is to avoid data loss, which can happen if an additional disk failure occurs in a RAID group that is already running in degraded mode, or if a stand-alone system encounters a catastrophic error and has to shut down without NVRAM. You can extend the number of hours the system operates in these conditions by increasing the value of this option (the default value is 24). You can even disable the shutdown by setting the option to 0, but the longer the system operates with one or both of these conditions, the greater the chance of incurring data loss. How Data ONTAP uses RAID to protect your data and data availability | 121
  • 122. How RAID-level disk scrubs verify data integrity RAID-level scrubbing means checking the disk blocks of all disks in use in aggregates (or in a particular aggregate, plex, or RAID group) for media errors and parity consistency. If Data ONTAP finds media errors or inconsistencies, it uses RAID to reconstruct the data from other disks and rewrites the data. RAID-level scrubs help improve data availability by uncovering and fixing media and checksum errors while the RAID group is in a normal state (for RAID-DP, RAID-level scrubs can also be performed when the RAID group has a single-disk failure). RAID-level scrubs can be scheduled or run manually. Changing the schedule for automatic RAID-level scrubs You can change the start time and duration of the scheduled scrubs if you want your data to be scrubbed more often than the default RAID-level scrub schedule allows. About this task The default RAID-level scrub schedule is to start a scrub every day starting at 1:00 a.m. On Sundays, the scrub runs for twelve hours; all other days it runs for four hours. Steps 1. Update the RAID-level scrub schedule: storage raid-options modify -node nodename -name raid.scrub.schedule - value duration[h|m]@weekday@start_time,[duration[h| m]@weekday@start_time,...] • duration is specified in either hours or minutes. • weekday is the day of the week: mon, tue, wed, thu, fri, sat, or sun. • start_time is the hour when the scrub should start, using 24-hour format. The current automatic RAID-level scrub schedule is replaced by the schedule you specify. 2. Confirm your new schedule: storage raid-options show raid.scrub.schedule Example The following command prevents any RAID-level scrubbing on node sys1-a from occurring on Mondays, and increases the time duration on Saturdays to eight hours: 122 | Physical Storage Management Guide
  • 123. storage raid-options modify -node sys1-a -name raid.scrub.schedule - value 4h@tue@1,4h@wed@1,4h@thu@1,4h@fri@1,8h@sat@1,12h@sun@1 How you run a manual RAID-level scrub You can manually run a RAID-level scrub on individual RAID groups, plexes, aggregates, or all aggregates using the storage aggregate scrub command. You can also stop, suspend, and resume manual RAID-level scrubs. If you try to run a RAID-level scrub on a RAID group that is not in a normal state (for example, a group that is reconstructing or degraded), the scrub returns errors and does not check that RAID group. You can run a RAID-level scrub on a RAID-DP group with one failed disk. Controlling the impact of RAID operations on system performance You can reduce the impact of RAID operations on system performance by decreasing the speed of the RAID operations. You can control the speed of the following RAID operations with RAID options: • RAID data reconstruction • Disk scrubbing • Plex resynchronization • Synchronous mirror verification The speed that you select for each of these operations might affect the overall performance of the storage system. However, if the operation is already running at the maximum speed possible and it is fully utilizing one of the three system resources (the CPU, disks, or the disk-to-controller connection bandwidth), changing the speed of the operation has no effect on the performance of the operation or the storage system. If the operation is not yet running, you can set a speed that minimally slows storage system network operations or a speed that severely slows storage system network operations. For each operation, use the following guidelines: • If you want to reduce the performance impact on client access to the storage system, change the specific RAID option from medium to low. Doing so also causes the operation to slow down. • If you want to speed up the operation, change the RAID option from medium to high. Doing so might decrease the performance of the storage system in response to client access. How Data ONTAP uses RAID to protect your data and data availability | 123
  • 124. Controlling the performance impact of RAID data reconstruction Because RAID data reconstruction consumes CPU resources, increasing the speed of data reconstruction sometimes slows storage system network and disk operations. You can control the speed of data reconstruction with the raid.reconstruc.perf_impact option. About this task When RAID data reconstruction and plex resynchronization are running at the same time, Data ONTAP limits the combined resource utilization to the greater impact set by either operation. For example, if raid.resync.perf_impact is set to medium and raid.reconstruct.perf_impact is set to low, the resource utilization of both operations has a medium impact. The setting for this option also controls the speed of Rapid RAID Recovery. Step 1. Enter the following command: storage raid-options modify -node node_name raid.reconstruct.perf_impact impact impact can be high, medium, or low. high means that the storage system uses most of the system resources available for RAID data reconstruction; this setting can heavily affect storage system performance, but reconstruction finishes sooner, reducing the time that the RAID group is in degraded mode. low means that the storage system uses very little of the system resources; this setting lightly affects storage system performance. However, reconstruction takes longer to complete, increasing the time that the storage system is running in degraded mode. The default impact is medium. Controlling the performance impact of RAID-level scrubbing When Data ONTAP performs a RAID-level scrub, it checks the disk blocks of all disks on the storage system for media errors and parity consistency. You can control the impact this operation has on system performance with the raid.verify.perf_impact option. About this task When RAID-level scrubbing and mirror verification are running at the same time, Data ONTAP limits the combined resource utilization to the greater impact set by either operation. For example, if raid.verify.perf_impact is set to medium and raid.scrub.perf_impact is set to low, the resource utilization by both operations has a medium impact. 124 | Physical Storage Management Guide
  • 125. If there are times during the day when the load on your storage system is decreased, you can also limit the performance impact of the automatic RAID-level scrub by changing the start time or duration of the automatic scrub. Step 1. Enter the following command: storage raid-options modify -node node_name raid.scrub.perf_impact impact impact can be high, medium, or low. high means that the storage system uses most of the system resources available for scrubbing; this setting can heavily affect storage system performance, but the scrub finishes sooner. low means that the storage system uses very little of the system resources; this setting lightly affects storage system performance, but the scrub takes longer to complete. The default impact is low. Controlling the performance impact of plex resynchronization Plex resynchronization ensures that both plexes of a mirrored aggregate are identical. You can control the performance impact of plex resynchronization by using the raid.resync.perf_impact option. About this task When plex resynchronization and RAID data reconstruction are running at the same time, Data ONTAP limits the combined resource utilization to the greatest impact set by either operation. For example, if raid.resync.perf_impact is set to medium and raid.reconstruct.perf_impact is set to low, the resource utilization by both operations has a medium impact. You should set this option to the same value for both nodes in an HA configuration. Step 1. Enter the following command: storage raid-options modify -node node_name raid.resync.perf_impact impact impact can be high, medium, or low. high means that the storage system uses most of the system resources available for plex resynchronization; this setting can heavily affect storage system performance, but the resynchronization finishes sooner. low means that the storage system uses very little of the system resources; this setting lightly affects storage system performance, but the resynchronization takes longer to complete. How Data ONTAP uses RAID to protect your data and data availability | 125
  • 126. The default impact is medium. Controlling the performance impact of mirror verification You use mirror verification to ensure that the two plexes of a synchronous mirrored aggregate are identical. You can control the speed of mirror verification, and its effect on system resources, by using the raid.verify.perf_impact option. About this task When mirror verification and RAID-level scrubbing are running at the same time, Data ONTAP limits the combined resource utilization to the greatest impact set by either operation. For example, if raid.verify.perf_impact is set to medium and raid.scrub.perf_impact is set to low, the resource utilization of both operations has a medium impact. For more information about synchronous mirroring, see the Clustered Data ONTAP Data Protection Tape Backup and Recovery Guide. Step 1. Enter the following command: storage raid-options modify -node node_name raid.verify.perf_impact impact impact can be high, medium, or low. high means that the storage system uses most of the system resources available for mirror verification; this setting can heavily affect storage system performance, but the mirror verification finishes sooner. low means that the storage system uses very little of the system resources; this setting lightly affects storage system performance, but the mirror verification takes longer to complete. The default impact is low. 126 | Physical Storage Management Guide
  • 127. What aggregates are To support the differing security, backup, performance, and data sharing needs of your users, you can group the physical data storage resources on your storage system into one or more aggregates. You can then design and configure these aggregates to provide the appropriate level of performance and redundancy. Each aggregate has its own RAID configuration, plex structure, and set of assigned drives or array LUNs. The aggregate provides storage, based on its configuration, to its associated FlexVol volumes or Infinite Volume. Aggregates have the following characteristics: • They can be composed of drives or array LUNs. • They can be mirrored or unmirrored. • If they are composed of drives, they can be single-tier (composed of only HDDs or only SSDs) or they can be Flash Pool aggregates, which include both HDD RAID groups and an SSD cache. The cluster administrator can assign one or more aggregates to a Storage Virtual Machine (SVM), in which case you can use only those aggregates to contain volumes for that SVM. Related information NetApp Technical Report 3437: Storage Subsystem Resiliency Guide How unmirrored aggregates work Unless you are using SyncMirror, all of your aggregates are unmirrored. Unmirrored aggregates have only one plex (copy of their data), which contains all of the RAID groups belonging to that aggregate. The following diagram shows an unmirrored aggregate composed of disks, with its one plex. The aggregate has four RAID groups: rg0, rg1, rg2, and rg3. Each RAID group has 6 data disks, one parity disk, and one dparity (double parity) disk. All disks used by the aggregate come from the same pool, pool0. 127
  • 128. Spare disk Data disk Parity disk dParity disk RAID group Aggregate Plex0 (pool0) pool0 rg0 rg1 rg2 rg3 Legend The following diagram shows an unmirrored aggregate with array LUNs, with its one plex. It has two RAID groups, rg0 and rg1. All array LUNs used by the aggregate come from the same pool, pool0. Aggregate Plex0 (pool0) rg0 rg1 Legend array LUN in the aggregate Data ONTAP RAID group 128 | Physical Storage Management Guide
  • 129. How mirrored aggregates work Mirrored aggregates have two plexes (copies of their data), which use the SyncMirror functionality to duplicate the data to provide redundancy. When a mirrored aggregate is created (or when a second plex is added to an existing unmirrored aggregate), Data ONTAP copies the data in the original plex (plex0) to the new plex (plex1). The plexes are physically separated (each plex has its own RAID groups and its own pool), and the plexes are updated simultaneously. This provides added protection against data loss if more disks fail than the RAID level of the aggregate protects against or there is a loss of connectivity, because the unaffected plex continues to serve data while you fix the cause of the failure. After the plex that had a problem is fixed, the two plexes resynchronize and reestablish the mirror relationship. Note: The time for the two plexes to resynchronize can vary and depends on many variables such as aggregate size, system load, how much data has changed, and so on. The disks and array LUNs on the system are divided into two pools: pool0 and pool1. Plex0 gets its storage from pool0 and plex1 gets its storage from pool1. The following diagram shows an aggregate composed of disks with SyncMirror enabled and implemented. A second plex has been created for the aggregate, plex1. The data in plex1 is a copy of the data in plex0, and the RAID groups are also identical. The 32 spare disks are allocated to pool0 or pool1, 16 disks for each pool. Aggregate Plex0 (pool0) Plex1 (pool1) pool0 pool1 rg0 rg1 rg2 rg3 rg0 rg1 rg2 rg3 Spare disk Data disk Parity disk dParity disk RAID group Legend What aggregates are | 129
  • 130. The following diagram shows an aggregate composed of array LUNs with SyncMirror enabled and implemented. A second plex has been created for the aggregate, plex1. Plex1 is a copy of plex0, and the RAID groups are also identical. Aggregate Plex0 (pool0) Plex1 (pool1) rg0rg0 rg1rg1 array LUN in the aggregate Data ONTAP RAID group What a Flash Pool aggregate is A Flash Pool aggregate combines both SSDs and HDDs (performance or capacity) to provide a high- performance aggregate more economically than an SSD aggregate. The SSDs provide a high-performance cache for the active data set of the data volumes provisioned on the Flash Pool aggregate, offloading random read operations and repetitive random write operations to improve response times and overall throughput for disk I/O-bound data access operations. (Performance is not significantly increased for predominately sequential workloads.) Related tasks Creating a Flash Pool aggregate using physical SSDs on page 171 How Flash Pool aggregates work In general, Flash Pool aggregates are used and managed in a similar manner as standard aggregates. However, you need to understand how both the SSD and HDD RAID groups interact and affect the rest of the system. The SSD RAID groups, also called the SSD cache, can be composed of physical SSDs or allocation units from SSD storage pools (but not both). The SSD cache does not contribute to the size of the aggregate as calculated against the maximum aggregate size. For example, even if an aggregate is at the maximum aggregate size, you can add an SSD RAID group to it. The SSDs do count toward the overall (node or HA pair) drive limit. 130 | Physical Storage Management Guide
  • 131. The HDD RAID groups in a Flash Pool aggregate behave the same as HDD RAID groups in a standard aggregate, following the same rules for mixing disk types, sizes, speeds, and checksums. For example, you cannot combine performance and capacity disks in the HDD RAID groups of a Flash Pool aggregate. The checksum type, RAID type, and RAID group size values can be configured for the SSD cache RAID groups and HDD RAID groups independently. If the Flash Pool aggregate uses an SSD storage pool for its SSD cache, the cache RAID type can be changed only when the first SSD RAID groups are added, and the size of the SSD RAID groups are determined by the number of SSDs in the storage pool. There is a platform-dependent maximum size for the SSD cache. Related concepts Rules for mixing drive types in Flash Pool aggregates on page 147 Rules for mixing HDD types in aggregates on page 146 How the available Flash Pool cache capacity is calculated on page 133 Related tasks Changing the RAID type of RAID groups in a Flash Pool aggregate on page 180 Related information NetApp Hardware Universe Requirements for using Flash Pool aggregates The Flash Pool technology has some configuration requirements that you should be aware of before planning to use it in your storage architecture. Flash Pool aggregates cannot be used in the following configurations: You cannot use the following types of storage objects in Flash Pool aggregates: • SSDs that are partitioned for root-data partitioning Partitioned HDDs can be used for the HDD RAID groups of a Flash Pool aggregate. • Array LUNs You can use Flash Pool aggregates and the Flash Cache module (WAFL external cache) in the same system. However, data stored in a Flash Pool aggregate is not cached in the Flash Cache module. Flash Cache is reserved for data stored in aggregates composed of only HDDs. You can use data compression on volumes associated with a Flash Pool aggregate. However, compressed blocks are cached in the Flash Pool cache only for read operations; compressed blocks are not cached for write operations. What aggregates are | 131
  • 132. Data in read-only volumes, such as SnapMirror or SnapVault destinations, is cached in the Flash Pool cache. Flash Pool aggregates can be created from mirrored aggregates; however, the SSD configuration must be kept the same for both plexes. If you create a Flash Pool aggregate using an aggregate that was created using Data ONTAP 7.1 or earlier, the volumes associated with that Flash Pool aggregate do not support write caching. Related information NetApp Hardware Universe NetApp Technical Report 4070: Flash Pool Design and Implementation Guide Considerations for RAID type and spare management for Flash Pool cache Because of the higher cost of SSDs, it is especially important to find the right balance between cost and protection against drive failure. Knowing how the performance and availability of Flash Pool aggregates are affected by SSDs becoming unavailable can help you determine the correct approach for your needs. The failure or unavailability of an SSD affects its RAID group and aggregate the same way as for HDDs. When an SSD being used in a Flash Pool cache becomes unavailable, the RAID group that contains that SSD goes into degraded mode, and the cache performance for that Flash Pool aggregate is reduced until the RAID group can be reconstructed by copying the contents of the unavailable SSD to a spare SSD. As with HDD RAID groups, if a cache RAID group experiences more concurrent SSD failures than its RAID level can correct for (3 for RAID-DP, 2 for RAID4 ), data integrity and availability can be compromised. Therefore, for maximum data availability, you should use RAID-DP for both the HDD RAID groups and the cache. In addition, you should have a spare SSD available at all times so that reconstruction can begin immediately in the case of an SSD failure. The spare SSD can be shared between multiple Flash Pool aggregates, although all Flash Pool aggregates using that spare will be without a spare after an SSD fails until the spare can be replaced. If using RAID-DP and providing a spare for the Flash Pool cache is prohibitively expensive, and if your RAID group size is not larger than the maximum RAID4 RAID group size, the next best alternative is to use RAID4 for the Flash Pool cache (you should still use RAID-DP for the HDD RAID groups), and provide a spare SSD. This is preferable to using RAID-DP with no spare, because double SSD failures are relatively rare, especially with a smaller RAID group size, and reconstruction takes less time than procuring and installing a spare. Note: If you are using SSD storage pools, you cannot change the RAID type of the cache RAID groups after they have been added to an aggregate. If you believe that you might want to increase the size of your storage pools over time, you should consider the maximum size you will require when determining the correct RAID type to use for the SSD cache. 132 | Physical Storage Management Guide
  • 133. How Flash Pool aggregates and Flash Cache compare Both the Flash Pool technology and the family of Flash Cache modules (Flash Cache and Flash Cache 2) provide a high-performance cache to increase storage performance. However, there are differences between the two technologies that you should understand before choosing between them. You can employ both technologies on the same system. However, data stored in volumes associated with a Flash Pool aggregate is not cached by Flash Cache. Criteria Flash Pool aggregate Flash Cache Scope A specific aggregate All aggregates assigned to a node Caching types supported Read and write Read Cached data availability during and after takeover events Cached data is available and unaffected by either planned or unplanned takeover events. Cached data is not available during takeover events. After giveback for a planned takeover, previously cached data that is still valid is re- cached automatically. PCIe slot on storage controller required? No Yes Compressed blocks cached? Yes No Supported with array LUNs? No Yes Supported with Storage Encryption? Yes Yes. Data in the cache is not encrypted. For more information about Flash Cache, see the Clustered Data ONTAP System Administration Guide for Cluster Administrators. How the available Flash Pool cache capacity is calculated Flash Pool cache capacity cannot exceed a platform-dependent limit for the node or HA configuration. Knowing the available cache capacity enables you to determine how much cache capacity you can add to your Flash Pool aggregates before reaching the limit. The current Flash Pool cache capacity is the sum of all of the Flash Pool caches (as reported by the storage aggregate show command). For nodes using SyncMirror, including MetroCluster configurations, only the Flash Pool cache of one plex counts toward the cache limit. If Flash Cache modules are installed in a node, the available cache capacity for Flash Pool use is the Flash Pool cache capacity limit minus the sum of the Flash Cache module cache installed on the node. (In the unusual case where the size of the Flash Cache modules is not symmetrical between the What aggregates are | 133
  • 134. two nodes in an HA configuration, the available Flash Pool cache capacity is decreased by the size of the larger Flash Cache module.) For nodes in an HA configuration, the cache size limit applies to the HA configuration as a whole, and can be split arbitrarily between the two nodes, provided that the total limit for the HA configuration is not exceeded. For nodes in a MetroCluster configuration, the cache size limit applies to the configuration as a whole, but it cannot be split arbitrarily between the nodes as it can for an HA configuration. This means that the limit for each node in the MetroCluster configuration is one quarter of the total cache size limit. If Flash Cache is present on any of the nodes, then the size of the largest Flash Cache module must be subtracted from the overall limit before the per-node limit is calculated. Calculation with Flash Cache modules For an HA configuration composed of two storage controllers with a maximum cache capacity of 12 TB and 2 TB of Flash Cache installed on each node, the maximum Flash Pool aggregate cache capacity for the HA pair would be 12 TB minus 2 TB, or 10 TB. Calculation with asymmetrically sized Flash Cache modules For an HA configuration composed of two storage controllers with a maximum cache capacity of 12 TB and 2 TB of Flash Cache installed on one node and 3 TB of Flash Cache installed on the other node, the maximum Flash Pool cache capacity for the HA pair would be 12 TB minus 3 TB, or 9 TB. Calculation for a MetroCluster configuration For a MetroCluster configuration composed of four storage controllers with a maximum cache capacity of 12 TB, the maximum Flash Pool cache capacity for each node would be 12 TB divided by four, or 3 TB. Calculation for a MetroCluster configuration with Flash Cache modules For a MetroCluster configuration composed of four storage contollers with a maximum cache capacity of 12 TB and 2 TB of Flash Cache installed on one node, the maximum Flash Pool cache capacity for each node would be 12 TB minus 2 TB divided by four, or 2.5 TB. Related concepts How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates on page 137 How Flash Pool aggregates work on page 130 134 | Physical Storage Management Guide
  • 135. Related references Commands for managing aggregates on page 193 Related information NetApp Hardware Universe Using caching policies on volumes in Flash Pool aggregates You can change the caching policy for a volume that resides on Flash Pool aggregates by using the - caching-policy parameter in the volume create command. When you create a volume on a Flash Pool aggregate, by default, the auto caching policy is assigned to the volume. In most cases, the default caching policy is preferable. The caching policy for a volume should be changed only if a different policy provides better performance. You can set the caching policies for a volume when you create a volume on the Flash Pool aggregate. You can modify the caching policy by using the volume modify command. The caching policies can also be moved between a Flash Pool aggregate and non-Flash Pool aggregate. The following is the list of caching policies, descriptions, and the combinations of read and write caching policies that you can set based on the usage of the volume: Policy Name Description Read caching policy Write caching policy Privilege auto Read caches all metadata blocks and randomly read user data blocks and write caches all randomly overwritten user data blocks. random_read random- write admin none Does not cache any user data or metadata blocks. none none admin random_read Read caches all metadata blocks and randomly read user data blocks. random_read none advanced noread- random_write Write caches all randomly overwritten user data blocks. none random- write advanced meta Read caches only metadata blocks. meta none advanced What aggregates are | 135
  • 136. Policy Name Description Read caching policy Write caching policy Privilege meta- random_write Read caches all metadata blocks and write caches all randomly overwritten user data blocks. meta random- write advanced random_read_ write Read caches all metadata, randomly read, and randomly written user data blocks. random_read _write none advanced random_read_ write-random- write Read caches all metadata, randomly read, and randomly written user data blocks. It also write caches randomly overwritten user data blocks. random_read _write random- write advanced Related tasks Determining and enabling volume write-caching eligibility for Flash Pool aggregates on page 178 Related information NetApp Technical Report 4070: Flash Pool Design and Implementation Guide Modifying caching policies You should modify the caching policy of a volume only if another policy is expected to provide better performance. The volume must be on a Flash Pool aggregate. Step 1. Use the volume modify command to change the caching policies of a volume. For information about the parameters that you can use with this command, see the volume modify man page. Example The following example modifies the caching policy of a volume named “vol2” to the policy none: volume modify -volume vol2 -caching-policy none 136 | Physical Storage Management Guide
  • 137. How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates Flash Pool SSD partitioning enables you to group SSDs together into an SSD storage pool that can be allocated to multiple Flash Pool aggregates. This amortizes the cost of the parity SSDs over more aggregates, increases SSD allocation flexibility, and maximizes SSD performance . The storage pool is associated with an HA pair, and can be composed of SSDs owned by either node in the HA pair. When you add an SSD to a storage pool, it becomes a shared SSD, and it is divided into 4 partitions. Storage from an SSD storage pool is divided into allocation units, each of which represents 25% of the total storage capacity of the storage pool. Allocation units contain one partition from each SSD in the storage pool, and are added to a Flash Pool cache as a single RAID group. By default, for storage pools associated with an HA pair, two allocation units are assigned to each of the HA partners, but you can reassign the allocation units to the other HA partner if needed (allocation units must be owned by the node that owns the aggregate). SSD storage pools do not have a RAID type. When an allocation unit is added to a Flash Pool aggregate, the appropriate number of partitions are designated to provide parity to that RAID group. The following diagram shows one example of Flash Pool SSD partitioning. The SSD storage pool pictured is providing cache to two Flash Pool aggregates: What aggregates are | 137
  • 138. Storage pool SP1 is composed of 5 SSDs; in addition, there is one hot spare SSD available to replace any SSD that experiences a failure. Two of the storage pool's allocation units are allocated to Flash Pool FP1, and two are allocated to Flash Pool FP2. FP1 has a cache RAID type of RAID4, so the allocation units provided to FP1 contain only one partition designated for parity. FP2 has a cache RAID type of RAID-DP, so the allocation units provided to FP2 include a parity partition and a double-parity partition. In this example, two allocation units are allocated to each Flash Pool aggregate; however, if one Flash Pool aggregate needed a larger cache, you could allocate three of the allocation units to that Flash Pool aggregate, and only one to the other. Related concepts What a Flash Pool aggregate is on page 130 Related tasks Creating a Flash Pool aggregate using SSD storage pools on page 173 Considerations for when to use SSD storage pools SSD storage pools provide many benefits, but they also introduce some restrictions that you should be aware of when deciding whether to use SSD storage pools or dedicated SSDs. SSD storage pools make sense only when they are providing cache to two or more Flash Pool aggregates. SSD storage pools provide the following benefits: • Increased storage utilization for SSDs used in Flash Pool aggregates SSD storage pools reduce the overall percentage of SSDs needed for parity by enabling you to share parity SSDs between two or more Flash Pool aggregates. • Ability to share spares between HA partners Because the storage pool is effectively owned by the HA pair, one spare, owned by one of the HA partners, can function as a spare for the entire SSD storage pool if needed. • Better utilization of SSD performance The high performance provided by SSDs can support access by both controllers in an HA pair. These advantages must be weighed against the costs of using SSD storage pools, which include the following items: • Reduced fault isolation The loss of a single SSD affects all RAID groups that include one of its partitions. In this situation, every Flash Pool aggregate that has cache allocated from the SSD storage pool that contains the affected SSD has one or more RAID groups in reconstruction. • Reduced performance isolation 138 | Physical Storage Management Guide
  • 139. If the Flash Pool cache is not properly sized, there can be contention for the cache between the Flash Pool aggregates that are sharing it. This risk can be mitigated with proper cache sizing and QoS controls. • Decreased management flexibility When you add storage to a storage pool, you increase the size of all Flash Pool caches that include one or more allocation units from that storage pool; you cannot determine how the extra capacity is distributed. How you use SSD storage pools To enable SSDs to be shared by multiple Flash Pool aggregates, you place them in a storage pool. After you add an SSD to a storage pool, you can no longer manage it as a stand-alone entity—you must use the storage pool to assign or allocate the storage provided by the SSD. You create storage pools for a specific HA pair. Then, you add allocation units from that storage pool to one or more Flash Pool aggregates owned by the same HA pair. Just as disks must be owned by the same node that owns an aggregate before they can be allocated to it, storage pools can provide storage only to Flash Pool aggregates owned by one of the nodes that owns the storage pool. If you need to increase the amount of Flash Pool cache on your system, you can add more SSDs to a storage pool, up to the maximum RAID group size for the RAID type of the Flash Pool caches using the storage pool. When you add an SSD to an existing storage pool, you increase the size of the storage pool's allocation units, including any allocation units that are already allocated to a Flash Pool aggregate. You should provide one or more spare SSDs for your storage pools, so that if an SSD in that storage pool becomes unavailable, Data ONTAP can use a spare SSD to reconstruct the partitions of the malfunctioning SSD. You do not need to reserve any allocation units as spare capacity; Data ONTAP can use only a full, unpartitioned SSD as a spare for SSDs in a storage pool. After you add an SSD to a storage pool, you cannot remove it, just as you cannot remove disks from an aggregate. If you want to use the SSDs in a storage pool as discrete drives again, you must destroy all Flash Pool aggregates to which the storage pool's allocation units have been allocated, and then destroy the storage pool. Requirements and best practices for using SSD storage pools There are some technologies that cannot be combined with Flash Pool aggregates that use SSD storage pools. You cannot use the following technologies with Flash Pool aggregates that use SSD storage pools for their cache storage: • MetroCluster • SyncMirror Mirrored aggregates can coexist with Flash Pool aggregates that use storage pools; however, Flash Pool aggregates cannot be mirrored. What aggregates are | 139
  • 140. • Physical SSDs Flash Pool aggregates can use SSD storage pools or physical SSDs, but not both. SSD storage pools must conform to the following rules: • SSD storage pools can contain only SSDs; HDDs cannot be added to an SSD storage pool. • SSD storage pools can contain between 3 and 28 SSDs. If an SSD storage pool contains more SSDs than the maximum RAID4 RAID group size for SSDs, then it cannot be used for a Flash Pool aggregate whose cache has a RAID type of RAID4. • All SSDs in an SSD storage pool must be owned by the same HA pair. • You cannot use SSDs that have been partitioned for root-data partitioning in a storage pool. If you provide storage from a single storage pool to two caches with different RAID types, and you expand the size of the storage pool beyond the maximum RAID group size for RAID4, the extra partitions in the RAID4 allocation units go unused. For this reason, it is a best practice to keep your cache RAID types homogenous for a storage pool. You cannot change the RAID type of cache RAID groups allocated from a storage pool. You set the RAID type for the cache before adding the first allocation units, and you cannot change it later. When you create a storage pool or add SSDs to an existing storage pool, you must use the same size SSDs. If a failure occurs and no spare of the correct size exists, Data ONTAP can use a larger SSD to replace the failed SSD. However, the larger SSD is right-sized to match the size of the other SSDs in the storage pool, resulting in lost SSD capacity. You can use only one spare SSD for a storage pool. If the storage pool provides allocation units to Flash Pool aggregates owned by both nodes in the HA pair, then the spare SSD can be owned by either node. However, if the storage pool provides allocation units only to Flash Pool aggregates owned by one of the nodes in the HA pair, then the SSD spare must be owned by that same node. Creating an SSD storage pool You create SSD storage pools to provide SSD cache for two to four Flash Pool aggregates. About this task Storage pools do not support a diskcount parameter; you must supply a disk list when creating or adding disks to a storage pool. Steps 1. Determine the names of the spare SSDs available to you: storage aggregate show-spare-disks -disk-type SSD The SSDs used in a storage pool can be owned by either node of an HA pair. 2. Create the storage pool: 140 | Physical Storage Management Guide
  • 141. storage pool create -storage-pool sp_name -disk-list disk1,disk2,... 3. Optional: Show the newly created storage pool: storage pool show -storage-pool sp_name Result After the SSDs are placed into the storage pool, they no longer appear as spares on the cluster, even though the storage provided by the storage pool has not yet been allocated to any Flash Pool caches. The SSDs can no longer be added to a RAID group as a discrete drive; their storage can be provisioned only by using the allocation units of the storage pool to which they belong. Related concepts How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates on page 137 Considerations for adding SSDs to an existing storage pool versus creating a new one on page 142 Related tasks Creating a Flash Pool aggregate using SSD storage pools on page 173 Adding SSDs to an SSD storage pool on page 141 Related references Commands for managing SSD storage pools on page 143 Adding SSDs to an SSD storage pool When you add SSDs to an SSD storage pool, you increase the storage pool's physical and usable sizes and allocation unit size. The larger allocation unit size also affects allocation units that have already been allocated to Flash Pool aggregates. Before you begin You must have determined that this operation will not cause you to exceed the cache limit for your HA pair. Data ONTAP does not prevent you from exceeding the cache limit when you add SSDs to an SSD storage pool, and doing so can render the newly added storage capacity unavailable for use. About this task When you add SSDs to an existing SSD storage pool, the SSDs must be owned by one node or the other of the same HA pair that already owned the existing SSDs in the storage pool. You can add SSDs that are owned by either node of the HA pair. Steps 1. Optional: View the current allocation unit size and available storage for the storage pool: What aggregates are | 141
  • 142. storage pool show -instance sp_name 2. Find available SSDs: storage disk show -container-type spare -type SSD 3. Add the SSDs to the storage pool: storage pool add -storage-pool sp_name -disk-list disk1,disk2… The system displays which Flash Pool aggregates will have their size increased by this operation and by how much, and prompts you to confirm the operation. Related concepts How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates on page 137 Considerations for adding SSDs to an existing storage pool versus creating a new one on page 142 How the available Flash Pool cache capacity is calculated on page 133 Related tasks Determining the impact to cache size of adding SSDs to an SSD storage pool on page 143 Creating an SSD storage pool on page 140 Related information NetApp Hardware Universe Considerations for adding SSDs to an existing storage pool versus creating a new one You can increase the size of your SSD cache in two ways—by adding SSDs to an existing SSD storage pool or by creating a new SSD storage pool. The best method for you depends on your configuration and plans for the storage. The choice between creating a new storage pool or adding storage capacity to an existing one is similar to deciding whether to create a new RAID group or add storage to an existing one: • If you are adding a large number of SSDs, creating a new storage pool provides more flexibility because you can allocate the new storage pool differently from the existing one. • If you are adding only a few SSDs, and increasing the RAID group size of your existing Flash Pool caches is not an issue, then adding SSDs to the existing storage pool keeps your spare and parity costs lower, and automatically allocates the new storage. If your storage pool is providing allocation units to Flash Pool aggregates whose caches have different RAID types, and you expand the size of the storage pool beyond the maximum RAID4 RAID group size, the newly added partitions in the RAID4 allocation units are unused. 142 | Physical Storage Management Guide
  • 143. Determining the impact to cache size of adding SSDs to an SSD storage pool If adding SSDs to a storage pool causes your platform model's cache limit to be exceeded, Data ONTAP does not allocate the newly added capacity to any Flash Pool aggregates. This can result in some or all of the newly added capacity being unavailable for use. About this task When you add SSDs to an SSD storage pool that has allocation units already allocated to Flash Pool aggregates, you increase the cache size of each of those aggregates and the total cache on the system. If none of the storage pool's allocation units have been allocated, adding SSDs to that storage pool does not affect the SSD cache size until one or more allocation units are allocated to a cache. Steps 1. Determine the usable size of the SSDs you are adding to the storage pool: storage disk show disk_name -fields usable-size 2. Determine how many allocation units remain unallocated for the storage pool: storage pool show-available-capacity sp_name All unallocated allocation units in the storage pool are displayed. 3. Calculate the amount of cache that will be added by applying the following formula: ( 4 – number of unallocated allocation units) × 25% × usable size × number of SSDs Commands for managing SSD storage pools Data ONTAP provides the storage pool command for managing SSD storage pools. If you want to... Use this command... Display how much storage a storage pool is providing to which aggregates storage pool show-aggregate Display how much cache would be added to the overall cache capacity for both RAID types (allocation unit data size) storage pool show -instance Display the disks in a storage pool storage pool show-disks Display the unallocated allocation units for a storage pool storage pool show-available- capacity What aggregates are | 143
  • 144. If you want to... Use this command... Change the ownership of one or more allocation units of a storage pool from one HA partner to the other storage pool reassign Related information Clustered Data ONTAP 8.3 Commands: Manual Page Reference How the SVM affects which aggregates can be associated with a FlexVol volume FlexVol volumes are always associated with one Storage Virtual Machine (SVM), and one aggregate that supplies its storage. The SVM can limit which aggregates can be associated with that volume, depending on how the SVM is configured. When you create a FlexVol volume, you specify which SVM the volume will be created on, and which aggregate that volume will get its storage from. All of the storage for the newly created FlexVol volume comes from that associated aggregate. If the SVM for that volume has aggregates assigned to it, then you can use only one of those assigned aggregates to provide storage to volumes on that SVM. This can help you ensure that your SVMs are not sharing physical storage resources inappropriately. This segregation can be important in a multi- tenancy environment, because for some space management configurations, volumes that share the same aggregate can affect each other's access to free space when space is constrained for the aggregate. Aggregate assignment requirements apply to both cluster administrators and SVM administrators. Volume move and volume copy operations are not constrained by the SVM aggregate assignments, so if you are trying to keep your SVMs on separate aggregates, you must ensure that you do not violate your SVM aggregate assignments when you perform those operations. If the SVM for that volume has no aggregates assigned to it, then the cluster administrator can use any aggregate in the cluster to provide storage to the new volume. However, the SVM administrator cannot create volumes for SVMs with no assigned aggregates. For this reason, if you want your SVM administrator to be able to create volumes for a specific SVM, then you must assign aggregates to that SVM. Changing the aggregates assigned to an SVM does not affect any existing volumes. For this reason, the list of aggregates assigned to an SVM cannot be used to determine the aggregates associated with volumes for that SVM. Related tasks Assigning aggregates to SVMs on page 190 144 | Physical Storage Management Guide
  • 145. Related information Clustered Data ONTAP 8.3 System Administration Guide for Cluster Administrators Understanding how Data ONTAP works with heterogeneous storage When you have disks with different characteristics (type, speed, size, checksum) or have both disks and array LUNs attached to your storage system, you have heterogeneous storage. Understanding how Data ONTAP works with heterogeneous storage helps you ensure that your aggregates and RAID groups follow best practices and provide maximum storage availability. How you can use disks with mixed speeds in the same aggregate Whenever possible, you should use disks of the same speed in an aggregate. However, if needed, you can configure Data ONTAP to allow mixed speed aggregates based on the disk class. To configure Data ONTAP to allow mixed speed aggregates, you use the following RAID options: • raid.mix.hdd.rpm.performance • raid.mix.hdd.rpm.capacity When these options are set to on, Data ONTAP allows mixing speeds for the designated disk class. Performance disk types are FC and SAS; capacity disk types are BSAS, FSAS, MSATA, and ATA. By default, raid.mix.hdd.rpm.performance is set to off, and raid.mix.hdd.rpm.capacity is set to on. Even if Data ONTAP is not configured to allow mixing speeds, you can still create aggregates out of disks with different speeds by setting the -allow-mixed parameter to true. How to control disk selection from heterogeneous storage When disks with different characteristics coexist on the same node, or when both disks and array LUNs are attached to the same node, the system has heterogeneous storage. When you create an aggregate from heterogeneous storage, you should take steps to ensure that Data ONTAP uses the disks you expect. If your node has heterogeneous storage and you do not explicitly specify what type of disks to use, Data ONTAP uses the disk type (including array LUNs) with the highest number of available disks. When you create or add storage to an aggregate using heterogeneous storage, you should use one of the following methods to ensure that Data ONTAP selects the correct disks or disk types: • Through disk attributes: ◦ You can specify disk size by using the -disksize option. Disks with a usable size between 90% and 105% of the specified size are selected. What aggregates are | 145
  • 146. ◦ You can specify disk speed by using the -diskrpm option. ◦ You can specify disk type by using the -disktype option. • Through disk selection preview. You can use the -simulate option to identify which disks Data ONTAP selects automatically. If you are happy with the disks selected, you can proceed with automatic disk selection. Otherwise, you can use one of the previous methods to ensure that the correct disks are selected. • Through an explicit disk list. You can list the names of specific disks you want to use. However, it is a best practice to use the diskcount option rather than disk lists; this allows Data ONTAP to make the best disk selection for your configuration. Note: For unplanned events such as disk failures, which cause Data ONTAP to add another disk to a RAID group automatically, the best way to ensure that Data ONTAP chooses the best disk for any RAID group on your system is to always have at least one spare (and preferably two) available to match all disk types and sizes in use in your system. Rules for mixing HDD types in aggregates You can mix disks from different loops or stacks within the same aggregate. Depending on the value of the raid.mix.hdd.disktype RAID options, you can mix certain types of HDDs within the same aggregate, but some disk type combinations are more desirable than others. When the appropriate raid.mix.hdd.disktype option is set to off, HDD RAID groups can be composed of only one Data ONTAP disk type. This setting ensures that your aggregates are homogeneous, and requires that you provide sufficient spare disks for every disk type in use in your system. The default value for the raid.mix.hdd.disktype.performance option is off, to prevent mixing SAS and FCAL disks. The default value for the raid.mix.hdd.disktype.capacity option is on. For this setting, the BSAS, FSAS, and ATA disk types are considered to be equivalent for the purposes of creating and adding to aggregates, and for spare management. To maximize aggregate performance and for easier storage administration, you should avoid mixing FC-connected and SAS-connected disks in the same aggregate. This is because of the performance mismatch between FC-connected storage shelves and SAS-connected storage shelves. When you mix these connection types in the same aggregate, the performance of the aggregate is limited by the presence of the FC-connected storage shelves, even though some of the data is being served from the higher-performing SAS-connected storage shelves. MSATA disks cannot be mixed with any other disk type in the same aggregate. If your node uses root-data partitioning, the same mixing rules apply to the partitioned HDDs as to HDDs that are not partitioned. Partitioned HDDs can be mixed with HDDs that are not partitioned in the same aggregate if they are otherwise compatible, but not the same RAID group. Physical 146 | Physical Storage Management Guide
  • 147. (unpartitioned) HDDs that are added to an existing RAID group that contains partitioned HDDs become partitioned. Disks using Storage Encryption have a Data ONTAP disk type of SAS. However, they cannot be mixed with any other disk type, including SAS disks that are not using Storage Encryption. If any disks on a storage system use Storage Encryption, all of the disks on the storage system (and its high- availability partner node) must use Storage Encryption. Note: If you set a raid.mix.hdd.disktype option to off for a system that already contains aggregates with more than one type of HDD, those aggregates continue to function normally and accept both types of HDDs. However, no other aggregates composed of the specified disk type will accept mixed HDD types as long as that option is set to off. Related concepts How Data ONTAP reports disk types on page 10 Related information NetApp Technical Report 3437: Storage Subsystem Resiliency Guide Rules for mixing drive types in Flash Pool aggregates By definition, Flash Pool aggregates contain more than one drive type. However, the HDD RAID groups follow the same drive-type mixing rules as standard aggregates. For example, you cannot mix performance and capacity disks in the same Flash Pool aggregate. The SSD cache can contain only SSDs. Rules for mixing storage in array LUN aggregates When planning for aggregates, you must consider the rules for mixing storage in aggregates. You cannot mix different storage types or array LUNs from different vendors or vendor families in the same aggregate. Adding the following to the same aggregate is not supported: • Array LUNs and disks • Array LUNs with different checksum types • Array LUNs from different drive types (for example, FC and SATA) or different speeds • Array LUNs from different storage array vendors • Array LUNs from different storage array model families Note: Storage arrays in the same family share the same performance and failover characteristics. For example, members of the same family all perform active-active failover, or they all perform active-passive failover. More than one factor might be used to determine storage array families. What aggregates are | 147
  • 148. For example, storage arrays with different architectures would be in different families even though other characteristics might be the same. How the checksum type is determined for array LUN aggregates Each Data ONTAP aggregate has a checksum type associated with it. The aggregate checksum type is determined by the checksum type of the array LUNs that are added to it. The checksum type of an aggregate is determined by the checksum type of the first array LUN that is added to the aggregate. The checksum type applies to an entire aggregate (that is, to all volumes in the aggregate). Mixing array LUNs of different checksum types in an aggregate is not supported. • An array LUN of type block must be used with block checksum type aggregates. • An array LUN of type zoned must be used with advanced zoned checksum (AZCS or advanced_zoned) type aggregates. Note: Prior to Data ONTAP 8.1.1, zoned checksum array LUNs were used with ZCS (zoned) type aggregates. Starting with 8.1.1, any new aggregates created with zoned checksum array LUNs are AZCS aggregates. However, you can add zoned checksum array LUNs to existing ZCS aggregates. Before you add array LUNs to an aggregate, you must know the checksum type of the LUNs you want to add, for the following reasons: • You cannot add array LUNs of different checksum types to the same aggregate. • You cannot convert an aggregate from one checksum type to the other. When you create an aggregate you can specify the number of array LUNs to be added, or you can specify the names of the LUNs to be added. If you want to specify a number of array LUNs to be added to the aggregate, the same number or more array LUNs of that checksum type must be available. How to determine space usage in an aggregate You can view space usage by all volumes in one or more aggregates with the aggregate show- space command. This helps you see which volumes are consuming the most space in their containing aggregates so that you can take actions to free more space. The used space in an aggregate is directly affected by the space used in the FlexVol volumes and Infinite Volume constituents it contains. Measures that you take to increase space in a volume also affect space in the aggregate. 148 | Physical Storage Management Guide
  • 149. When the aggregate is offline, no values are displayed. Only non-zero values are displayed in the command output. However, you can use the -instance parameter to display all possible feature rows regardless of whether they are enabled and using any space. A value of - indicates that there is no data available to display. The following rows are included in the aggregate show-space command output: • Volume Footprints The total of all volume footprints within the aggregate. It includes all of the space that is used or reserved by all data and metadata of all volumes in the containing aggregate. It is also the amount of space that is freed if all volumes in the containing aggregate are destroyed. Infinite Volume constituents appear in the output of space usage commands as if the constituents were FlexVol volumes. • Aggregate Metadata The total file system metadata required by the aggregate, such as allocation bitmaps and inode files. • Snapshot Reserve The amount of space reserved for aggregate Snapshot copies, based on volume size. It is considered used space and is not available to volume or aggregate data or metadata. • Snapshot Reserve Unusable The amount of space originally allocated for aggregate Snapshot reserve that is unavailable for aggregate Snapshot copies because it is being used by volumes associated with the aggregate. Can occur only for aggregates with a non-zero aggregate Snapshot reserve. • Total Used The sum of all space used or reserved in the aggregate by volumes, metadata, or Snapshot copies. • Total Physical Used The amount of space being used for data now (rather than being reserved for future use). Includes space used by aggregate Snapshot copies. There is never a row for Snapshot spill. The following example shows the aggregate show-space command output for an aggregate whose Snapshot reserve is 5%. If the Snapshot reserve was 0, the row would not be displayed. cluster1::> storage aggregate show-space Aggregate : wqa_gx106_aggr1 Feature Used Used% -------------------------------- ---------- ------ Volume Footprints 101.0MB 0% Aggregate Metadata 300KB 0% Snapshot Reserve 5.98GB 5% What aggregates are | 149
  • 150. Total Used 6.07GB 5% Total Physical Used 34.82KB 0% How you can determine and control a volume's space usage in the aggregate You can determine which FlexVol volumes and Infinite Volume constituents are using the most space in the aggregate and specifically which features within the volume. The volume show- footprint command provides information about a volume's footprint, or its space usage within the containing aggregate. The volume show-footprint command shows details about the space usage of each volume in an aggregate, including offline volumes. This command does not directly correspond to the output of the df command, but instead bridges the gap between the output of the volume show-space and aggregate show-space commands. All percentages are calculated as a percent of aggregate size. Only non-zero values are displayed in the command output. However, you can use the -instance parameter to display all possible feature rows regardless of whether they are enabled and using any space. A value of - indicates that there is no data available to display. Infinite Volume constituents appear in the output of space usage commands as if the constituents were FlexVol volumes. The following example shows the volume show-footprint command output for a volume called testvol: cluster1::> volume show-footprint testvol Vserver : thevs Volume : testvol Feature Used Used% -------------------------------- ---------- ----- Volume Data Footprint 120.6MB 4% Volume Guarantee 1.88GB 71% Flexible Volume Metadata 11.38MB 0% Delayed Frees 1.36MB 0% Total Footprint 2.01GB 76% The following table explains some of the key rows of the output of the volume show-footprint command and what you can do to try to decrease space usage by that feature: 150 | Physical Storage Management Guide
  • 151. Row/feature name Description/contents of row Some ways to decrease Volume Data Footprint The total amount of space used in the containing aggregate by a volume's data in the active file system and the space used by the volume's Snapshot copies. This row does not include reserved space, so if volumes have reserved files, the volume's total used space in the volume show-space command output can exceed the value in this row. • Deleting data from the volume. • Deleting Snapshot copies from the volume. Volume Guarantee The amount of space reserved by the volume in the aggregate for future writes. The amount of space reserved depends on the guarantee type of the volume. Changing the type of guarantee for the volume to none. This row will go to 0. If you configure your volumes with a volume guarantee of none, you should refer to Technical Report 3965 or 3483 for information about how a volume guarantee of none can affect storage availability. Flexible Volume Metadata The total amount of space used in the aggregate by the volume's metadata files. No direct method to control. Delayed Frees Blocks that Data ONTAP used for performance and cannot be immediately freed. When Data ONTAP frees blocks in a FlexVol volume, this space is not always immediately shown as free in the aggregate because operations to free the space in the aggregate are batched for increased performance. Blocks that are declared free in the FlexVol volume but that are not yet free in the aggregate are called “delayed free blocks” until the associated delayed free blocks are processed. For SnapMirror destinations, this row has a value of 0 and is not displayed. No direct method to control. What aggregates are | 151
  • 152. Row/feature name Description/contents of row Some ways to decrease File Operation Metadata The total amount of space reserved for file operation metadata. After space is used for file operation metadata, it is not returned as free space to the aggregate, but it is reused by subsequent file operations. No direct method to control. Total Footprint The total amount of space that the volume uses in the aggregate. It is the sum of all of the rows. Any of the methods used to decrease space used by a volume. How Infinite Volumes use aggregates Each Infinite Volume distributes data across multiple aggregates from multiple nodes. By understanding the way that Infinite Volumes use aggregates, you can plan your aggregates in a way that supports the Infinite Volumes that you want. For more information about Infinite Volumes, see the Clustered Data ONTAP Infinite Volumes Management Guide. Aggregate requirements for Infinite Volumes The aggregates that are used by an Infinite Volume should be larger than 100 TB with a minimum of 1.1 TB of available space. If the Infinite Volume uses storage classes, the aggregates must also meet the requirements of the storage class. If an aggregate has less than 1.1 TB of available space, it is not used by the Storage Virtual Machine (SVM) with Infinite Volume. If the Infinite Volume uses storage classes, aggregates must meet the requirements of the storage class to be used. For example, if the storage class is designated to use aggregates of type SAS, aggregates created for that storage class must consist entirely of SAS disks. How FlexVol volumes and Infinite Volumes share aggregates Aggregates can be shared among the volumes in a cluster. Each aggregate can contain multiple FlexVol volumes alongside multiple constituents of Infinite Volumes. When you create an Infinite Volume, constituents of the Infinite Volume are placed on aggregates that are assigned to its containing Storage Virtual Machine (SVM). If the SVM with Infinite Volume includes aggregates that contain FlexVol volumes, one or more of the Infinite Volume's constituents might be placed on aggregates that already include FlexVol volumes, if those aggregates meet the requirements for hosting Infinite Volumes. 152 | Physical Storage Management Guide
  • 153. Similarly, when you create a FlexVol volume, you can associate that FlexVol volume with an aggregate that is already being used by an Infinite Volume. The following diagram illustrates aggregate sharing in a four-node cluster that includes both FlexVol volumes and an Infinite Volume. The Infinite Volume uses the aggregates aggrA, aggrB, aggrC, aggrD, aggrE, and aggrG even though the aggregates aggrB, aggrC, and aggrG already provide storage to FlexVol volumes. (For clarity, the individual constituents that make up the Infinite Volume are not shown.) Infinite Volume aggrA aggrC aggrB aggrD aggrE aggrF aggrHaggrG Node 1 Node 2 Node 3 Node 4 FV FV FV FV FV FV FV FV FV FV FV How storage classes affect which aggregates can be associated with Infinite Volumes Each storage class definition specifies an aggregate type. When you create an Infinite Volume with a storage class, only the type of aggregate specified for the storage class can supply storage for the volume. You must understand storage class definitions to create aggregates that are appropriate for the storage class. Storage class definitions are available only in OnCommand Workflow Automation. After you understand the aggregate requirements for each storage class, you can use the command-line interface or OnCommand Workflow Automation to create aggregates for storage classes. However, you must use OnCommand Workflow Automation, not the command-line interface, to create an Infinite Volume with one or more storage classes. When you use OnCommand Workflow Automation to create an Infinite Volume with a storage class, OnCommand Workflow Automation automatically filters the aggregates available in the cluster based on the storage class that you want to use. If no aggregates meet the requirements of the storage class, you cannot create an Infinite Volume with that storage class. What aggregates are | 153
  • 154. How aggregates and nodes are associated with Infinite Volumes The aggregate list of the containing Storage Virtual Machine (SVM) with Infinite Volume determines which aggregates the Infinite Volume uses, as well as who can create an Infinite Volume and which nodes the Infinite Volume uses. The aggregate list can be specified or unspecified, which is represented as a dash ("-"). By default, when a cluster administrator creates any SVM, its aggregate list is unspecified. After the SVM is created, the cluster administrator can specify the aggregate list by using the vserver add- aggregates command. Considerations when choosing to specify the aggregate list or leave it unspecified If you are dedicating an entire cluster to the SVM with Infinite Volume, you can leave the aggregate list of an SVM with Infinite Volume unspecified. In most other situations, you should specify the aggregate list of an SVM with Infinite Volume. Leaving the aggregate list of an SVM with Infinite Volume unspecified has the following outcomes: • Only a cluster administrator can create the Infinite Volume, not an SVM administrator. • When the Infinite Volume is created, it uses all nodes in the cluster. • When the Infinite Volume is created, it can potentially use all of the aggregates in the cluster. How the aggregate list contains candidate aggregates The aggregate list of an SVM with Infinite Volume acts only as a candidate aggregate list for an Infinite Volume. An Infinite Volume uses aggregates according to various factors, including the following requirements: • When an Infinite Volume is created, at least one data constituent is created on at least one aggregate from each node in the aggregate list. • An Infinite Volume uses only the aggregates that it requires to meet the capacity requirements for its specified size. If the assigned aggregates have far greater capacity than the Infinite Volume requires when it is first created, some aggregates in the aggregate list might not contain any Infinite Volume constituents. How the aggregate list determines the nodes An Infinite Volume uses every node that has an aggregate in the aggregate list of an SVM with Infinite Volume. 154 | Physical Storage Management Guide
  • 155. When changes to the aggregate list take effect Changes to the aggregate list do not have any immediate effect. The aggregate list is used only when the size of an Infinite Volume changes. For example, if you add an aggregate to the aggregate list of an SVM with Infinite Volume, that aggregate is not used until you modify the size of the Infinite Volume. If you add aggregates from a new node to the aggregate list and then resize the Infinite Volume, whether the Infinite Volume uses the aggregates from the new node depends on several variables, including the size of existing constituents and how much the Infinite Volume was increased in size. How the aggregate list can be filtered You can filter the aggregate list for the SVM by using advanced parameters that control which aggregates are used for each type of constituent, such as data constituents. Unlike the aggregate list for the SVM, these aggregate-selection parameters apply only to a single operation. For example, if you use the parameter for data constituent aggregates when you create the Infinite Volume and then resize the Infinite Volume without using the parameter, the Infinite Volume uses the SVM aggregate list. How space is allocated inside a new Infinite Volume Several rules govern how space is allocated to constituents when an Infinite Volume is created. Understanding these rules can help you understand the best practices for configuring aggregates for an Infinite Volume. The following rules govern how space is allocated to constituents when an Infinite Volume is created: 1. The namespace constituent and its mirror copies are created. Before any space is allocated for data, the namespace constituent and namespace mirror constituents are created as big as possible within their maximum sizes. a. The namespace constituent is placed on the aggregate with the most available space on any node that the Infinite Volume uses. b. The first namespace mirror constituent is placed on the aggregate with the next most available space, as long as the aggregate is on a node that meets all of the following conditions: • The node is used by the Infinite Volume. • It does not already contain the namespace constituent. • It is preferably not the partner node in the HA pair of the node that contains the namespace constituent. c. If SnapDiff is enabled, additional namespace mirror constituents are placed on the aggregate with the most available space on each remaining node used by the Infinite Volume. 2. The data capacity is divided equally among the nodes that the Infinite Volume uses. What aggregates are | 155
  • 156. The data capacity of an Infinite Volume is balanced across nodes. Data capacity is the space remaining from the Infinite Volume's size after deducting the space required by the namespace- related constituents. 3. Within each node, individual data constituents are made as big as possible within a specified maximum. Data constituents are always created as big as they are allowed to be within a specified maximum. Each time that Data ONTAP creates a data constituent, it evaluates all of the aggregates that the Infinite Volume uses on the node and selects the aggregate that has the most available space. Relocating ownership of aggregates used by Infinite Volumes If you want to relocate ownership of aggregates that are used by an Infinite Volume with SnapDiff enabled, you must ensure that the destination node has a namespace mirror constituent, and you must perform the aggregate reallocation in a specific order. Before you begin • You must know whether SnapDiff is enabled on the Infinite Volume. If you do not know, you can use the volume show command with the -fields -enable- snapdiff parameter. • You must know the names of the aggregates on the source and destination nodes. About this task • Follow this procedure only if SnapDiff is enabled on an Infinite Volume that uses the node containing the aggregates that are affected by the ownership change. If SnapDiff is not enabled on the Infinite Volume, do not follow this procedure, because the presence of an Infinite Volume does not affect the aggregate relocation operation. • For more information about SnapDiff and about how Infinite Volumes use aggregates, see the Clustered Data ONTAP Infinite Volumes Management Guide. Steps 1. Determine whether the destination node is already used by the Infinite Volume by using the vserver show command with the -instance parameter. • If the Aggregate List includes an aggregate from the destination node, the Infinite Volume already uses the destination node. • If the Aggregate List does not include an aggregate from the destination node, the Infinite Volume does not currently use the destination node. The destination node must have a namespace mirror constituent before you can relocate aggregate ownership. 156 | Physical Storage Management Guide
  • 157. 2. If the destination node is not currently used by the Infinite Volume, add the destination node to the Infinite Volume. a. Determine the size of the Infinite Volume's namespace constituent by using the volume show command with the -is-constituent true parameter and identifying the size of the constituents with “ns” in their names. b. Identify an aggregate on the destination node that has available space to accommodate a namespace mirror constituent by using the aggregate show command. c. Assign the aggregate from the new node to the Infinite Volume by using the vserver modify command with the -aggr-list parameter. When you specify the aggregate list, you should include all the existing aggregates from existing nodes as well as the new aggregate from the new node. d. Increase the size of the Infinite Volume by using the volume modify command with the - size parameter. Increase the size of the Infinite Volume by an amount that is equal or larger than the size of the namespace constituent. A namespace mirror constituent is created on the destination node. 3. Identify the type of Infinite Volume constituents contained by each aggregate on the source node by using the volume show command with the -vserver and -is-constituent true parameters. Constituents with “data” in their names are data constituents. The constituent with a name ending in “_ns” is the namespace constituent. Constituents with “ns_mirror” in their names are namespace mirror constituents. Example In the following output, aggr3 contains a data constituent, aggr1 contains a namespace mirror constituent, and aggr2 contains a namespace mirror constituent: cluster1::> volume show -vserver vs0 -is-constituent true Vserver Volume Aggregate State Type Size Available Used% --------- ------------ ------------ ---------- ---- ---------- ---------- ----- vs0 repo_vol_1024_data0001 aggr3 online RW 100TB 95TB 5% vs0 repo_vol_1024_data0002 vs_aggr online RW 100TB 95TB 5% vs0 repo_vol_1024_data0003 aggr4 online RW 100TB 95TB 5% ... ... ... vs0 repo_vol_ns aggr1 online RW 10TB 9.5TB 5% vs0 repo_vol_ns_mirror0001 aggr2 online DP 10TB 9.5TB 5% 100 entries were displayed. 4. Perform the aggregate relocation in the following way: What aggregates are | 157
  • 158. a. Divide the aggregates into the following two categories: • The single aggregate that contains either a namespace constituent or a namespace mirror constituent. It might also contain data constituents. • All the other aggregates on the node that contain data constituents. b. Relocate ownership of all aggregates that do not contain a namespace constituent or namespace mirror constituent. Note: If you want to relocate ownership of the aggregate that contains the namespace constituent without also relocating ownership of all of the aggregates that contain data constituents, you must contact technical support to create a namespace mirror constituent on the source node. c. If the remaining aggregate contains only the namespace constituent or if the remaining aggregate contains more than one type of constituent, relocate ownership of the remaining aggregate. If the remaining aggregate contains only a namespace mirror constituent and no data constituents, you do not need to change ownership of the aggregate. After you finish You can consider contacting technical support to delete any excess namespace mirror constituents that are using space unnecessarily. 158 | Physical Storage Management Guide
  • 159. Managing aggregates You create and manage your aggregates so that they can provide storage to their associated volumes. Creating an aggregate using unpartitioned drives You create an aggregate to provide storage to one or more FlexVol volumes and Infinite Volumes. Aggregates are a physical storage object; they are associated with a specific node in the cluster. Before you begin • You should know what drives or array LUNs will be used in the new aggregate. • You should have determined the correct RAID group size given the number of drives or array LUNs you are using to create the aggregate and your plans for expanding the aggregate in the future. • If you have multiple drive types in your node (heterogeneous storage), you should understand how you can ensure that the correct drive type is selected. About this task This procedure should not be used to create an aggregate composed of root or data partitions. Drives and array LUNs are owned by a specific node; when you create an aggregate, all drives in that aggregate must be owned by the same node, which becomes the home node for that aggregate. Aggregate names must conform to the following requirements: • They must begin with either a letter or an underscore (_). • They can contain only letters, digits, and underscores. • They can contain 250 or fewer characters. Steps 1. Display a list of available spares: storage aggregate show-spare-disks -original-owner node_name 2. Create the aggregate by using the storage aggregate create command. If you are logged in to the cluster on the cluster management interface, you can create an aggregate on any node in the cluster. To ensure that the aggregate is created on a specific node, use the -node parameter or specify drives that are owned by that node. 159
  • 160. The following list describes some of the options you can specify: • Aggregate's home node (that is, the node on which the aggregate is located unless the aggregate fails over to the node's storage failover partner) • The number of disks or array LUNs to be added, or a list of specific drives or array LUNs that are to be included in the aggregate For optimal disk selection, use the diskcount parameter rather than a disk list. This allows Data ONTAP to make an optimal disk selection for your configuration. • Checksum style to use for the aggregate • Type of drives to use • Size of drives to use • Drive speed to use • RAID type for RAID groups on the aggregate • Maximum number of drives or array LUNs that can be included in a RAID group • Whether drives with different RPM are allowed 3. Monitor zeroing progress and verify the RAID groups and drives of your new aggregate: storage aggregate show-status aggr_name Example You want to create a storage configuration for a single-node cluster and have 4 shelves of SAS disks totaling 96 drives, with 3 drives used for the root aggregate. You decide to allocate the remaining drives in the following manner: • 2 spares • 91 drives used for data in one aggregate: ◦ 4 RAID groups of 18 drives each ◦ 1 RAID group of 19 drives To accomplish this goal, you use the following steps: 1. Create the first RAID group of 19 disks: storage aggregate create -aggregate n01sas01 -node cl-01 -diskcount 19 -maxraidsize 19 -disktype SAS 2. Change the aggregate's maximum RAID group size to 18: storage aggregate modify -aggregate n01sas01 -maxraidsize 18 160 | Physical Storage Management Guide
  • 161. 3. Add the remaining 54 disks as 4 RAID groups of 18 disks each: storage aggregate add-disks -aggregate n01sas01 -node cl-01 - diskcount 54 -disktype SAS Later, if you added a 24-drive shelf to the node and wanted to add the storage to this aggregate, you could again modify the RAID group size to 24 so that the entire shelf could be accommodated in a single RAID group, provided that a RAID group size of 24 is acceptable for your installation. Related concepts What aggregates are on page 127 Considerations for sizing RAID groups on page 114 Understanding how Data ONTAP works with heterogeneous storage on page 145 Related tasks Creating an aggregate using root-data partitioning on page 161 Creating a Flash Pool aggregate using physical SSDs on page 171 Creating an aggregate using root-data partitioning Typically, root aggregates are created in the factory, and data aggregates are created during initial system setup. However, if you need to create a new data aggregate using partitioned drives, there are some differences in the procedure from creating an aggregate using physical, unpartitioned drives. Before you begin You should know what drives or partitions will be used in the new aggregate. You should have determined the correct RAID group size given the number of drives or partitions you are using to create the aggregate and your plans for expanding the aggregate in the future. About this task Drives and partitions are owned by a specific node; when you create an aggregate, all of the partitions you use for the new aggregate must be owned by the same node, which becomes the home node for the new aggregate. When you provision partitions, you must ensure that you do not leave the node without a disk with both partitions as spare. If you do, and the node experiences a controller disruption, valuable information about the problem (the core file) might not be available to provide to technical support. Aggregate names must conform to the following requirements: • They must begin with either a letter or an underscore (_). Managing aggregates | 161
  • 162. • They can contain only letters, digits, and underscores. • They can contain 250 or fewer characters. Steps 1. View the list of spare data partitions: storage aggregate show-spare-disks -original-owner node_name The list of disks with at least one spare partition is displayed. Data partitions are shown under Local Data Usable. You must use data partitions when you create a data aggregate. 2. Determine how many partitions you want to use in the aggregate. Remember to leave an appropriate number of spare data partitions. Data ONTAP will not use a root partition as a spare for a data aggregate. 3. Simulate the creation of the aggregate: storage aggregate create -aggregate aggr_name -node node_name -diskcount number_of_partitions -simulate true This enables you to see the result of the aggregate creation without actually provisioning any storage. If any warnings are displayed from the simulated command, you can adjust the command and repeat the simulation. 4. Create the aggregate: storage aggregate create -aggregate aggr_name -node node_name -diskcount number_of_partitions 5. Check zeroing status and verify the composition of the aggregate: storage aggregate show-status aggr_name Related concepts Understanding root-data partitioning on page 30 Considerations for sizing RAID groups on page 114 Related tasks Creating an aggregate using unpartitioned drives on page 159 Correcting misaligned spare partitions on page 169 162 | Physical Storage Management Guide
  • 163. Increasing the size of an aggregate that uses physical drives You can add disks or array LUNs to an aggregate so that it can provide more storage to its associated volumes. Before you begin • You must understand the requirement to add disks or array LUNs owned by the same system and pool • For aggregates composed of disks, you must understand the following: ◦ Benefits of keeping your RAID groups homogeneous for disk size and speed ◦ Which types of disks can be used together ◦ Checksum rules when disks of more than one checksum type are in use ◦ How to ensure that the correct disks are added to the aggregate (the disk addition operation cannot be undone) ◦ How to add disks to aggregates from heterogeneous storage ◦ Minimum number of disks to add for best performance ◦ Number of hot spares you need to provide for protection against disk failures ◦ Requirements for adding disks from multi-disk carrier disk shelves ◦ Requirement to add storage to both plexes of a mirrored aggregate at the same time to ensure that the plexes are the same size and contain the same disk types ◦ If you are adding cache to a Flash Pool aggregate, the cache limit for your system model and how much cache you are adding towards the limit About this task This procedure should not be used for aggregates composed of root or data partitions. Following these best practices when you add storage to an aggregate optimizes aggregate performance: • Add a complete RAID group at one time. The new RAID group does not have to be exactly the same size as the existing RAID groups, but it should not be less than one half the size of the existing RAID groups. Managing aggregates | 163
  • 164. • If any small RAID groups exist already, you can bring them up to the size of the other RAID groups, as long as you add at least as many data drives as are already in the RAID group. • Avoid adding a small number of drives to an existing RAID group. Doing so results in the added disks being the target for a disproportionate percentage of new data, causing the new disks to become a performance bottleneck. Steps 1. Verify that appropriate spare disks or array LUNs are available for you to add: storage aggregate show-spare-disks -original-owner node_name For disks, make sure that enough of the spares listed are of the correct type, size, speed, and checksum type for the target RAID group in the aggregate to which you are adding the disks. 2. Add the disks or array LUNs: storage aggregate add-disks -aggregate aggr_name [-raidgroup raid_group_name] disks If you are adding disks with a different checksum than the aggregate, as when creating a Flash Pool aggregate, or if you are adding disks to a mixed checksum aggregate, you must use the - checksumstyle parameter. If you are adding disks to a Flash Pool aggregate, you must use the -disktype parameter to specify the disk type. If you specify the -raidgroup parameter, the storage is added to the RAID group you specify. raid_group_name is the name that Data ONTAP gave to the group—for example, rg0. If you are adding SSDs to the SSD cache of a Flash Pool aggregate, you do not need to specify the RAID group name; the SSD RAID group is selected by default based on the type of the disks you are adding. disks specifies the disks to be added in one of the following ways: • -diskcount, usually further qualified by disk type or checksum type • -disklist disk1 [disk2...] If possible, you should use the diskcount option. Doing so allows Data ONTAP to optimize the disk selection for your configuration. If you are adding disks to a mirrored aggregate and you are specifying disk names, you must also use the -mirror-disklist parameter. Related concepts How to control disk selection from heterogeneous storage on page 145 Considerations for sizing RAID groups on page 114 What a Flash Pool aggregate is on page 130 164 | Physical Storage Management Guide
  • 165. Rules for mixing drive types in Flash Pool aggregates on page 147 Rules for mixing HDD types in aggregates on page 146 How the available Flash Pool cache capacity is calculated on page 133 What happens when you add storage to an aggregate on page 170 Related tasks Increasing the size of an aggregate that uses root-data partitioning on page 165 Related information NetApp Technical Report 4070: Flash Pool Design and Implementation Guide Increasing the size of an aggregate that uses root-data partitioning When you add storage to an existing aggregate that is using partitioned drives, you should be aware of whether you are adding a partitioned drive or an unpartitioned drive, and the tradeoffs of mixing those types of drives in the same RAID group versus creating a new RAID group. Before you begin • You should understand whether you are adding partitions or unpartitioned drives to the aggregate. • You should know what the RAID group size is for the aggregate you are adding the storage to. About this task When you add storage to an existing aggregate, you can let Data ONTAP choose the RAID group to add the storage to, or you can designate the target RAID group for the added storage, including creating a new RAID group. If an unpartitioned drive is added to a RAID group composed of partitioned drives, the new drive is partitioned, leaving an unused spare partition. If you do not want the new drive to be partitioned, you can add it to a RAID group that contains only unpartitioned (physical) drives. However, partitioning a drive might be preferable to creating a new, small RAID group. Following these best practices when you add storage to an aggregate optimizes aggregate performance: • Add a complete RAID group at one time. The new RAID group does not have to be exactly the same size as the existing RAID groups, but it should not be less than one half the size of the existing RAID groups. • If any small RAID groups exist already, you can bring them up to the size of the other RAID groups, as long as you add at least as many data drives as are already in the RAID group. Managing aggregates | 165
  • 166. • Avoid adding a small number of drives to an existing RAID group. Doing so results in the added drives being the target for a disproportionate percentage of new data, causing the new drives to become a performance bottleneck. When you provision partitions, you must ensure that you do not leave the node without a drive with both partitions as spare. If you do, and the node experiences a controller disruption, valuable information about the problem (the core file) might not be available to provide to technical support. Steps 1. Show the available spare storage on the system that owns the aggregate: storage aggregate show-spare-disks -original-owner node_name You can use the -is-disk-shared parameter to show only partitioned drives or only unpartitioned drives. 2. Show the current RAID groups for the aggregate: storage aggregate show-status aggr_name 3. Simulate adding the storage to the aggregate: storage aggregate add-disks -aggregate aggr_name -diskcount number_of_disks_or_partitions -simulate true This enables you to see the result of the storage addition without actually provisioning any storage. If any warnings are displayed from the simulated command, you can adjust the command and repeat the simulation. 4. Add the storage to the aggregate: storage aggregate add-disks -aggregate aggr_name -diskcount number_of_disks_or_partitions You can use the -raidgroup parameter if you want to add the storage to a different RAID group than the default. If you are adding partitions to the aggregate, you must use a disk that shows available capacity for the required partition type. For example, if you are adding partitions to a data aggregate (and using a disk list), the disk names you use must show available capacity in the Local Data Usable column. 5. Verify that the storage was added successfully: storage aggregate show-status -aggregate aggr_name 6. Ensure that the node still has at least one drive with both the root partition and the data partition as spare: storage aggregate show-spare-disks -original-owner node_name 166 | Physical Storage Management Guide
  • 167. If the node does not have a drive with both partitions as spare and it experiences a controller disruption, then valuable information about the problem (the core file) might not be available to provide to technical support. Example: Adding partitioned drives to an aggregate The following example shows that the cl1-s2 node has multiple spare partitions available: cl1-s2::> storage aggregate show-spare-disks -original-owner cl1-s2 -is-disk-shared true Original Owner: cl1-s2 Pool0 Shared HDD Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size Status --------------------------- ----- ------ -------------- -------- -------- -------- -------- 1.0.1 BSAS 7200 block 753.8GB 73.89GB 828.0GB zeroed 1.0.2 BSAS 7200 block 753.8GB 0B 828.0GB zeroed 1.0.3 BSAS 7200 block 753.8GB 0B 828.0GB zeroed 1.0.4 BSAS 7200 block 753.8GB 0B 828.0GB zeroed 1.0.8 BSAS 7200 block 753.8GB 0B 828.0GB zeroed 1.0.9 BSAS 7200 block 753.8GB 0B 828.0GB zeroed 1.0.10 BSAS 7200 block 0B 73.89GB 828.0GB zeroed 2 entries were displayed. The following example shows that the data_1 aggregate is composed of a single RAID group of five partitions: cl1-s2::> storage aggregate show-status -aggregate data_1 Owner Node: cl1-s2 Aggregate: data_1 (online, raid_dp) (block checksums) Plex: /data_1/plex0 (online, normal, active, pool0) RAID Group /data_1/plex0/rg0 (normal, block checksums) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ---------- shared 1.0.10 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.5 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.6 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.11 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.0 0 BSAS 7200 753.8GB 828.0GB (normal) 5 entries were displayed. The following example shows which partitions would be added to the aggregate: cl1-s2::> storage aggregate add-disks data_1 -diskcount 5 -simulate true Addition of disks would succeed for aggregate "data_1" on node "cl1-s2". The following disks would be used to add to the aggregate: 1.0.2, 1.0.3, 1.0.4, 1.0.8, 1.0.9. The following example adds five spare data partitions to the aggregate: cl1-s2::> storage aggregate add-disks data_1 -diskcount 5 The following example shows that the data partitions were successfully added to the aggregate: cl1-s2::> storage aggregate show-status -aggregate data_1 Owner Node: cl1-s2 Aggregate: data_1 (online, raid_dp) (block checksums) Plex: /data_1/plex0 (online, normal, active, pool0) RAID Group /data_1/plex0/rg0 (normal, block checksums) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ---------- Managing aggregates | 167
  • 168. shared 1.0.10 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.5 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.6 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.11 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.0 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.2 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.3 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.4 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.8 0 BSAS 7200 753.8GB 828.0GB (normal) shared 1.0.9 0 BSAS 7200 753.8GB 828.0GB (normal) 10 entries were displayed. The following example verifies that an entire disk, disk 1.0.1, remains available as a spare: cl1-s2::> storage aggregate show-spare-disks -original-owner cl1-s2 -is-disk-shared true Original Owner: cl1-s2 Pool0 Shared HDD Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size Status --------------------------- ----- ------ -------------- -------- -------- -------- -------- 1.0.1 BSAS 7200 block 753.8GB 73.89GB 828.0GB zeroed 1.0.10 BSAS 7200 block 0B 73.89GB 828.0GB zeroed 2 entries were displayed. Related concepts How to control disk selection from heterogeneous storage on page 145 What a Flash Pool aggregate is on page 130 Considerations for sizing RAID groups on page 114 Rules for mixing drive types in Flash Pool aggregates on page 147 Rules for mixing HDD types in aggregates on page 146 How the available Flash Pool cache capacity is calculated on page 133 What happens when you add storage to an aggregate on page 170 Related tasks Correcting misaligned spare partitions on page 169 Increasing the size of an aggregate that uses physical drives on page 163 Related information NetApp Technical Report 4070: Flash Pool Design and Implementation Guide 168 | Physical Storage Management Guide
  • 169. Correcting misaligned spare partitions When you add partitioned disks to an aggregate, you must leave a disk with both the root and data partition available as spare for every node. If you do not and your node experiences a disruption, Data ONTAP might not be able to create a core file. Before you begin You must have both a spare data partition and a spare root partition on the same type of disk owned by the same node. Steps 1. Display the spare partitions for the node: storage aggregate show-spare-disks -original-owner node_name Note which disk has a spare data partition (spare_data) and which disk has a spare root partition (spare_root). The spare partition will show a non-zero value under the Local Data Usable or Local Root Usable column. 2. Replace the disk with a spare data partition with the disk with the spare root partition: storage disk replace -disk spare_data -replacement spare_root -action start You can copy the data in either direction; however, copying the root partition takes less time to complete. 3. Monitor the progress of the disk replacement: storage aggregate show-status -aggregate aggr_name 4. After the replacement operation is complete, display the spares again to confirm that you have a full spare disk: storage aggregate show-spare-disks -original-owner node_name You should see a spare disk with usable space under both Local Data Usable and Local Root Usable. Example You display your spare partitions for node c1-01 and see that your spare partitions are not aligned: c1::> storage aggregate show-spare-disks -original-owner c1-01 Original Owner: c1-01 Pool0 Shared HDD Spares Local Local Data Root Physical Managing aggregates | 169
  • 170. Disk Type RPM Checksum Usable Usable Size --------------------------- ----- ------ -------------- -------- -------- -------- 1.0.1 BSAS 7200 block 753.8GB 0B 828.0GB 1.0.10 BSAS 7200 block 0B 73.89GB 828.0GB You start the disk replacement job: c1::> storage disk replace -disk 1.0.1 -replacement 1.0.10 -action start While you are waiting for the replacement operation to finish, you display the progress of the operation: c1::> storage aggregate show-status -aggregate aggr0_1 Owner Node: c1-01 Aggregate: aggr0_1 (online, raid_dp) (block checksums) Plex: /aggr0_1/plex0 (online, normal, active, pool0) RAID Group /aggr0_1/plex0/rg0 (normal, block checksums) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ---------- shared 1.0.1 0 BSAS 7200 73.89GB 828.0GB (replacing, copy in progress) shared 1.0.10 0 BSAS 7200 73.89GB 828.0GB (copy 63% completed) shared 1.0.0 0 BSAS 7200 73.89GB 828.0GB (normal) shared 1.0.11 0 BSAS 7200 73.89GB 828.0GB (normal) shared 1.0.6 0 BSAS 7200 73.89GB 828.0GB (normal) shared 1.0.5 0 BSAS 7200 73.89GB 828.0GB (normal) After the replacement operation is complete, you confirm that you have a full spare disk: ie2220::> storage aggregate show-spare-disks -original-owner c1-01 Original Owner: c1-01 Pool0 Shared HDD Spares Local Local Data Root Physical Disk Type RPM Checksum Usable Usable Size --------------------------- ----- ------ -------------- -------- -------- -------- 1.0.1 BSAS 7200 block 753.8GB 73.89GB 828.0GB What happens when you add storage to an aggregate By default, Data ONTAP adds new drives or array LUNs to the most recently created RAID group until it reaches its maximum size. Then Data ONTAP creates a new RAID group. Alternatively, you can specify a RAID group that you want to add storage to. When you create an aggregate or add storage to an aggregate, Data ONTAP creates new RAID groups as each RAID group is filled with its maximum number of drives or array LUNs. The last RAID group formed might contain fewer drives or array LUNs than the maximum RAID group size for the aggregate. In that case, any storage added to the aggregate is added to the last RAID group until the specified RAID group size is reached. If you increase the RAID group size for an aggregate, new drives or array LUNs are added only to the most recently created RAID group; the previously created RAID groups remain at their current size unless you explicitly add storage to them. 170 | Physical Storage Management Guide
  • 171. If you add a drive to a RAID group that is larger than the drives already there, the new drive is capacity-limited to be the same size as the other drives. Note: You are advised to keep your RAID groups homogeneous when possible. If needed, you can replace a mismatched drive with a more suitable drive later. Related tasks Increasing the size of an aggregate that uses physical drives on page 163 Increasing the size of an aggregate that uses root-data partitioning on page 165 Creating a Flash Pool aggregate using physical SSDs You create a Flash Pool aggregate by enabling the feature on an existing aggregate composed of HDD RAID groups, and then adding one or more SSD RAID groups to that aggregate. This results in two sets of RAID groups for that aggregate: SSD RAID groups (the SSD cache) and HDD RAID groups. Before you begin • You must have identified a valid aggregate composed of HDDs to convert to a Flash Pool aggregate. • You must have determined write-caching eligibility of the volumes associated with the aggregate, and completed any required steps to resolve eligibility issues. • You must have determined the SSDs you will be adding, and these SSDs must be owned by the node on which you are creating the Flash Pool aggregate. • You must have determined the checksum types of both the SSDs you are adding and the HDDs already in the aggregate. • You must have determined the number of SSDs you are adding and the optimal RAID group size for the SSD RAID groups. Using fewer RAID groups in the SSD cache reduces the number of parity disks required, but larger RAID groups require RAID-DP. • You must have determined the RAID level you want to use for the SSD cache. • You must have determined the maximum cache size for your system and determined that adding SSD cache to your aggregate will not cause you to exceed it. • You must have familiarized yourself with the configuration requirements for Flash Pool aggregates. Managing aggregates | 171
  • 172. About this task After you add an SSD cache to an aggregate to create a Flash Pool aggregate, you cannot remove the SSD cache to convert the aggregate back to its original configuration. By default, the RAID level of the SSD cache is the same as the RAID level of the HDD RAID groups. You can override this default selection by specifying the raidtype option when you add the first SSD RAID groups. Steps 1. Mark the aggregate as eligible to become a Flash Pool aggregate: storage aggregate modify -aggregate aggr_name -hybrid-enabled true If this step does not succeed, determine write-caching eligibility for the target aggregate. 2. Add the SSDs to the aggregate by using the storage aggregate add command. You can specify the SSDs by ID or by using the diskcount and disktype parameters. If the HDDs and the SSDs do not have the same checksum type, or if the aggregate is a mixed- checksum aggregate, then you must use the checksumstyle parameter to specify the checksum type of the disks you are adding to the aggregate. You can specify a different RAID type for the SSD cache by using the raidtype parameter. If you want the cache RAID group size to be different from the default for the RAID type you are using, you should change it now, by using the -cache-raid-group-size parameter. Related concepts What a Flash Pool aggregate is on page 130 Requirements for using Flash Pool aggregates on page 131 Considerations for RAID type and spare management for Flash Pool cache on page 132 How the available Flash Pool cache capacity is calculated on page 133 Related tasks Creating a Flash Pool aggregate using SSD storage pools on page 173 Determining Flash Pool candidacy and optimal cache size on page 175 Determining and enabling volume write-caching eligibility for Flash Pool aggregates on page 178 Creating an aggregate using unpartitioned drives on page 159 Related information NetApp Technical Report 4070: Flash Pool Design and Implementation Guide 172 | Physical Storage Management Guide
  • 173. Creating a Flash Pool aggregate using SSD storage pools You create a Flash Pool aggregate with SSD storage pools by enabling the feature on an existing aggregate composed of HDD RAID groups, and then adding one or more SSD storage pool allocation units to that aggregate. Before you begin • You must have identified a valid aggregate composed of HDDs to convert to a Flash Pool aggregate. • You must have determined write-caching eligibility of the volumes associated with the aggregate, and completed any required steps to resolve eligibility issues. • You must have created an SSD storage pool to provide the SSD cache to this Flash Pool aggregate. Any allocation unit from the storage pool that you want to use must be owned by the same node that owns the Flash Pool aggregate. • You must have determined how much cache you want to add to the aggregate. You add cache to the aggregate by allocation units. You can increase the size of the allocation units later by adding SSDs to the storage pool if there is room. • You must have determined the RAID type you want to use for the SSD cache. After you add a cache to the aggregate from SSD storage pools, you cannot change the RAID type of the cache RAID groups. • You must have determined the maximum cache size for your system and determined that adding SSD cache to your aggregate will not cause you to exceed it. You can see the amount of cache that will be added to the total cache size by using the storage pool show command. • You must have familiarized yourself with the configuration requirements for Flash Pool aggregates. About this task If you want the RAID type of the cache to different from that of the HDD RAID groups, you must specify the cache RAID type when you add the SSD capacity. After you add the SSD capacity to the aggregate, you can no longer change the RAID type of the cache. After you add an SSD cache to an aggregate to create a Flash Pool aggregate, you cannot remove the SSD cache to convert the aggregate back to its original configuration. Managing aggregates | 173
  • 174. Steps 1. Mark the aggregate as eligible to become a Flash Pool aggregate: storage aggregate modify -aggregate aggr_name -hybrid-enabled true If this step does not succeed, determine write-caching eligibility for the target aggregate. 2. Show the available SSD storage pool allocation units: storage pool show-available-capacity 3. Add the SSD capacity to the aggregate: storage aggregate add aggr_name -storage-pool sp_name -allocation-units number_of_units If you want the RAID type of the cache to be different from that of the HDD RAID groups, you must change it when you enter this command by using the raidtype parameter. You do not need to specify a new RAID group; Data ONTAP automatically puts the SSD cache into separate RAID groups from the HDD RAID groups. You cannot set the RAID group size of the cache; it is determined by the number of SSDs in the storage pool. The cache is added to the aggregate and the aggregate is now a Flash Pool aggregate. Each allocation unit added to the aggregate becomes its own RAID group. 4. Optional: Confirm the presence and size of the SSD cache: storage aggregate show aggr_name The size of the cache is listed under Total Hybrid Cache Size. Related concepts How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates on page 137 What a Flash Pool aggregate is on page 130 Requirements for using Flash Pool aggregates on page 131 Considerations for RAID type and spare management for Flash Pool cache on page 132 How the available Flash Pool cache capacity is calculated on page 133 Related tasks Creating a Flash Pool aggregate using physical SSDs on page 171 Creating an aggregate using unpartitioned drives on page 159 Determining Flash Pool candidacy and optimal cache size on page 175 Determining and enabling volume write-caching eligibility for Flash Pool aggregates on page 178 174 | Physical Storage Management Guide
  • 175. Related information NetApp Technical Report 4070: Flash Pool Design and Implementation Guide Determining Flash Pool candidacy and optimal cache size Before converting an existing aggregate to a Flash Pool aggregate, you can determine whether the aggregate is I/O bound, and what would be the best Flash Pool cache size for your workload and budget. You can also check whether the cache of an existing Flash Pool aggregate is sized correctly. Before you begin You should know approximately when the aggregate you are analyzing experiences its peak load. Steps 1. Enter advanced mode: set advanced 2. If you need to determine whether an existing aggregate would be a good candidate for conversion to a Flash Pool aggregate, determine how busy the disks in the aggregate are during a period of peak load, and how that is affecting latency: statistics show-periodic -object disk:raid_group -instance raid_group_name -counter disk_busy|user_read_latency -interval 1 - iterations 60 You can decide whether reducing latency by adding Flash Pool cache makes sense for this aggregate. Example The following command shows the statistics for the first RAID group of the aggregate “aggr1”: statistics show-periodic -object disk:raid_group -instance /aggr1/ plex0/rg0 -counter disk_busy|user_read_latency -interval 1 -iterations 60 3. Start AWA: system node run -node node_name wafl awa start aggr_name AWA begins collecting workload data for the volumes associated with the specified aggregate. 4. Exit advanced mode: set admin 5. Allow AWA to run until one or more intervals of peak load have occurred. Managing aggregates | 175
  • 176. AWA analyzes data for up to one rolling week in duration. Running AWA for more than one week will report only on data collected from the previous week. Cache size estimates are based on the highest loads seen during the data collection period. You do not need to ensure that the load is high for the entire data collection period. AWA collects workload statistics for the volumes associated with the specified aggregate. 6. Enter advanced mode: set advanced 7. Display the workload analysis: system node run -node node_name wafl awa print AWA displays the workload statistics and optimal Flash Pool cache size. 8. Stop AWA: system node run -node node_name wafl awa stop All workload data is flushed and is no longer available for analysis. 9. Exit advanced mode: set admin Example In the following example, AWA was run on aggregate “aggr1”. Here is the output of the awa print command after AWA had been running for about 3 days (442 10-minute intervals): ### FP AWA Stats ### Basic Information Aggregate aggr1 Current-time Mon Jul 28 16:02:21 CEST 2014 Start-time Thu Jul 31 12:07:07 CEST 2014 Total runtime (sec) 264682 Interval length (sec) 600 Total intervals 442 In-core Intervals 1024 Summary of the past 442 intervals max Read Throughput 39.695 MB/s Write Throughput 17.581 MB/s Cacheable Read (%) 92 % Cacheable Write (%) 83 % Max Projected Cache Size 114 GiB Projected Read Offload 82 % Projected Write Offload 82 % 176 | Physical Storage Management Guide
  • 177. Summary Cache Hit Rate vs. Cache Size Size 20% 40% 60% 80% 100% Read Hit 34 51 66 75 82 Write Hit 35 44 53 62 82 The entire results and output of Automated Workload Analyzer (AWA) are estimates. The format, syntax, CLI, results and output of AWA may change in future Data ONTAP releases. AWA reports the projected cache size in capacity. It does not make recommendations regarding the number of data SSDs required. Please follow the guidelines for configuring and deploying Flash Pool; that are provided in tools and collateral documents. These include verifying the platform cache size maximums and minimum number and maximum number of data SSDs. ### FP AWA Stats End ### ________________________________________ The results provide the following pieces of information: • Read Throughput and Write Throughput The throughput measurements can help you identify an aggregate that is receiving a higher amount of traffic. Note that these numbers do not indicate whether that aggregate is I/O bound. • Max Projected Cache Size The size at which the SSD cache would hold every eligible data block that was requested from disk during the AWA run. Note that this does not guarantee a hit for all future I/O operations, because they might request data that is not in the cache. However, if the workload during the AWA run was a typical one, and if your budget allows for it, this would be an ideal size for your Flash Pool cache. • Projected Read Offload and Projected Write Offload The approximate percentages of read and write operations that would have been handled by a Flash Pool cache of the optimal size rather than going to disk (projected cache hit rate). Note that this number is related to the performance increase you would see by converting the aggregate to a Flash Pool aggregate, but not an exact prediction. • Summary Cache Hit Rate vs. Cache Size This table can help you predict the performance impact of decreasing the size of the SSD cache from Max Projected Cache Size. These values are highly impacted by your workload. Depending on whether data that was aged out of the cache was ever accessed again, the impact of decreasing the size of the cache might be large or almost nonexistent. Managing aggregates | 177
  • 178. You can use this table to find the “sweet spot” of cost versus performance for your workload and budget. Related tasks Determining and enabling volume write-caching eligibility for Flash Pool aggregates on page 178 Related information NetApp Technical Report 4070: Flash Pool Design and Implementation Guide Determining and enabling volume write-caching eligibility for Flash Pool aggregates Understanding whether the FlexVol volumes associated with an aggregate are eligible for write caching can help you ensure that the volumes with high performance requirements can get the maximum performance improvement from having their associated aggregate converted to a Flash Pool aggregate. About this task Flash Pool aggregates employ two types of caching: read caching and write caching. Read caching is available for all volumes. Write caching is available for most volumes, but might be disabled for some volumes due to an internal ID collision. You determine write caching eligibility to help you decide which aggregates are good candidates to become Flash Pool aggregates. You do not need any SSDs to complete this procedure. Steps 1. Attempt to enable the Flash Pool capability on the aggregate: storage aggregate modify aggr_name -hybrid-enabled true 2. Take the applicable action based on the result of Step 1: If... Then... The Flash Pool capability is successfully enabled Disable the Flash Pool capability again: storage aggregate modify aggr_name -hybrid- enabled false You have completed this task. All of the volumes in the aggregate are eligible for write caching. 178 | Physical Storage Management Guide
  • 179. If... Then... Data ONTAP displays an error message telling you that the aggregate cannot be converted to a Flash Pool aggregate Determine which volumes are not eligible: volume show -volume * -fields hybrid-cache-write- caching-ineligibility-reason -aggregate aggr_name Each volume in the aggregate is listed, along with its reason for ineligibility if it is ineligible. Eligible volumes display a hyphen (“-”). 3. Your next steps depend on your requirements for the ineligible volumes: If you... Then... Do not need write caching enabled on the ineligible volumes You have completed this task. You must use the -force-hybrid- enabled option when you convert the aggregate to a Flash Pool aggregate. Need write caching enabled on the ineligible volumes You must move (or copy and delete) all but one of each set of volumes with the same conflicting ID to another aggregate and then move them back until no more volumes show an ID conflict. Example with ID collisions The following example shows the system output when there are ID collisions: clus1::> vol show -volume * -fields hybrid-cache-write-caching- ineligibility-reason -aggregate aggr1 (volume show) vserver volume hybrid-cache-write-caching-ineligibility-reason ------- -------- ----------------------------------------------- vs0 root_vs0 - vs0 vol1 - vs0 vol2 "ID Collision(27216)" vs0 vol3 "ID Collision(27216)" 4 entries were displayed. Related tasks Determining Flash Pool candidacy and optimal cache size on page 175 Managing aggregates | 179
  • 180. Changing the RAID type of RAID groups in a Flash Pool aggregate The SSD cache of a Flash Pool aggregate can have a different RAID type than the HDD RAID groups if you need to reduce the parity overhead for your SSD RAID groups. Before you begin • You should understand the considerations for RAID type and spare management for Flash Pool cache. • If you are changing the RAID type of the SSD cache from RAID-DP to RAID4, you should ensure that doing so will not cause you to exceed your cache size limit. About this task You can change the RAID type of the SSD cache for a Flash Pool aggregate using SSD storage pools only when you add the first storage pool allocation units to the aggregate. If the SSD cache has a different RAID type than the HDD RAID groups, the Flash Pool aggregate is considered to have a mixed RAID type, displayed as mixed_raid_type for the aggregate. In this case, the RAID type is also displayed for each RAID group. All HDD RAID groups in a Flash Pool aggregate must have the same RAID type, which should be RAID-DP. Steps 1. Change the RAID type of the SSD cache or HDD RAID groups of the Flash Pool aggregate: storage aggregate modify -aggregate aggr_name -raidtype raid_type - disktype disk_type To change the RAID type of the SSD cache, use -disktype SSD. To change the RAID type of the HDD RAID groups, specify any disk type included in the HDD RAID groups. 2. Verify the RAID groups in your Flash Pool aggregate: storage aggregate show -aggregate aggr_name You also can use the storage aggregate show-status command to obtain more details about the RAID types of the HDD RAID groups and SSD cache of the Flash Pool aggregate. Example In this example, the HDD RAID groups and SSD cache of a Flash Pool aggregate using physical SSDs named “test” initially have a RAID type of RAID-DP. The following command 180 | Physical Storage Management Guide
  • 181. changes the RAID type of the SSD cache to RAID4, and converts the Flash Pool aggregate to the mixed RAID type: storage aggregate modify -aggregate test -raidtype raid4 -disktype SSD The output from the storage aggregate show-status command shows that the aggregate has a mixed RAID type, the HDD RAID groups have a RAID type of RAID-DP, and the SSD cache has a RAID type of RAID4. storage aggregate show-status test Aggregate test (online, mixed_raid_type, hybrid) (block checksums) Plex /test/plex0 (online, normal, active, pool0) RAID Group /test/plex0/rg0 (normal, block checksums, raid-dp) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ---------- dparity 1.2.3 0 BSAS 7200 827.7GB 828.0GB (normal) parity 1.2.4 0 BSAS 7200 827.7GB 828.0GB (normal) data 1.2.5 0 BSAS 7200 827.7GB 828.0GB (normal) data 1.2.6 0 BSAS 7200 827.7GB 828.0GB (normal) data 1.2.8 0 BSAS 7200 827.7GB 828.0GB (normal) RAID Group /test/plex0/rg1 (normal, block checksums, raid4) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ---------- parity 1.3.3 0 SSD - 82.59GB 82.81GB (normal) data 1.4.0 0 SSD - 82.59GB 82.81GB (normal) data 1.4.1 0 SSD - 82.59GB 82.81GB (normal) data 1.4.2 0 SSD - 82.59GB 82.81GB (normal) Related concepts Considerations for RAID type and spare management for Flash Pool cache on page 132 Related information NetApp Hardware Universe Determining drive and RAID group information for an aggregate Some aggregate administration tasks require that you know what types of drives compose the aggregate, their size, checksum, and status, whether they are shared with other aggregates, and the size and composition of the RAID groups. Step 1. Show the drives for the aggregate, by RAID group: Managing aggregates | 181
  • 182. storage aggregate show-status aggr_name The drives are displayed for each RAID group in the aggregate. You can see the RAID type of the drive (data, parity, dparity) in the Position column. If the Position column displays shared, then the drive is shared: if it is an HDD, it is a partitioned disk; if it is an SSD, it is part of a storage pool. Example: a Flash Pool aggregate using an SSD storage pool and data partitions cluster1::> storage aggregate show-status nodeA_fp_1 Owner Node: cluster1-a Aggregate: nodeA_fp_1 (online, mixed_raid_type, hybrid) (block checksums) Plex: /nodeA_fp_1/plex0 (online, normal, active, pool0) RAID Group /nodeA_fp_1/plex0/rg0 (normal, block checksums, raid_dp) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ------- shared 2.0.1 0 SAS 10000 472.9GB 547.1GB (normal) shared 2.0.3 0 SAS 10000 472.9GB 547.1GB (normal) shared 2.0.5 0 SAS 10000 472.9GB 547.1GB (normal) shared 2.0.7 0 SAS 10000 472.9GB 547.1GB (normal) shared 2.0.9 0 SAS 10000 472.9GB 547.1GB (normal) shared 2.0.11 0 SAS 10000 472.9GB 547.1GB (normal) RAID Group /nodeA_flashpool_1/plex0/rg1 (normal, block checksums, raid4) (Storage Pool: SmallSP) Usable Physical Position Disk Pool Type RPM Size Size Status -------- --------------------------- ---- ----- ------ -------- -------- ------- shared 2.0.13 0 SSD - 186.2GB 745.2GB (normal) shared 2.0.12 0 SSD - 186.2GB 745.2GB (normal) 8 entries were displayed. Related concepts Understanding root-data partitioning on page 30 How Flash Pool SSD partitioning increases cache allocation flexibility for Flash Pool aggregates on page 137 Relocating aggregate ownership within an HA pair You can change the ownership of aggregates among the nodes in an HA pair without interrupting service from the aggregates. Both nodes in an HA pair are physically connected to each other's disks or array LUNs. Each disk or array LUN is owned by one of the nodes. While ownership of disks temporarily changes when a takeover occurs, the aggregate relocation operations either permanently (for example, if done for load balancing) or temporarily (for example, if done as part of takeover) change the ownership of all disks or array LUNs within an aggregate from one node to the other. The ownership changes without any data-copy processes or physical movement of the disks or array LUNs. 182 | Physical Storage Management Guide
  • 183. How aggregate relocation works Aggregate relocation takes advantage of the HA configuration to move the ownership of storage aggregates within the HA pair. Aggregate relocation enables storage management flexibility not only by optimizing performance during failover events, but also facilitating system operational and maintenance capabilities that previously required controller failover. Aggregate relocation occurs automatically during manually initiated takeovers to reduce downtime during planned failover events such as nondisruptive software upgrades. You can manually initiate aggregate relocation independent of failover for performance load balancing, system maintenance, and nondisruptive controller upgrades. However, you cannot use the aggregate relocation operation to move ownership of the root aggregate. The following illustration shows the relocation of the ownership of aggregate aggr_1 from Node1 to Node2 in the HA pair: Node2Node1 Aggregate aggr_1 8 disks on shelf sas_1 (shaded grey) Owned by Node1 before relocation Owned by Node2 after relocation The aggregate relocation operation can relocate the ownership of one or more SFO aggregates if the destination node can support the number of volumes in the aggregates. There is only a brief interruption of access to each aggregate. Ownership information is changed one by one for the aggregates. During takeover, aggregate relocation happens automatically after you manually initiate takeover. Before the target controller is taken over, ownership of each of the controller's aggregates is moved, one at a time, to the partner controller. When giveback is initiated, ownership is automatically moved back to the original node. The ‑bypass‑optimization parameter can be used with the storage failover takeover command to suppress aggregate relocation during the takeover. Managing aggregates | 183
  • 184. Aggregate relocation and Infinite Volumes with SnapDiff enabled The aggregate relocation requires additional steps if the aggregate is currently used by an Infinite Volume with SnapDiff enabled. You must ensure that the destination node has a namespace mirror constituent, and make decisions about relocating aggregates that include namespace constituents. Clustered Data ONTAP 8.3 Infinite Volumes Management Guide Related tasks Relocating ownership of aggregates used by Infinite Volumes on page 156 How root-data partitioning affects aggregate relocation If you have a platform model that uses root-data partitioning, also called shared disks, aggregate relocation processing occurs just as with physical (nonshared) disks. The container disk ownership changes to the destination node during aggregate relocation only if the operation transfers ownership of all partitions on that physical disk to the destination node. This ownership change occurs only with permanent aggregate relocation operations. Ownership changes that occur during negotiated storage failover takeover or giveback events are temporary. Relocating aggregate ownership You can change the ownership of an aggregate only between the nodes within an HA pair. About this task • Because volume count limits are validated programmatically during aggregate relocation operations, it is not necessary to check for this manually. If the volume count exceeds the supported limit, the aggregate relocation operation fails with a relevant error message. • You should not initiate aggregate relocation when system-level operations are in progress on either the source or the destination node; likewise, you should not start these operations during the aggregate relocation. These operations can include the following: ◦ Takeover ◦ Giveback ◦ Shutdown ◦ Another aggregate relocation operation ◦ Disk ownership changes 184 | Physical Storage Management Guide
  • 185. ◦ Aggregate or volume configuration operations ◦ Storage controller replacement ◦ Data ONTAP upgrade ◦ Data ONTAP revert • If you have a MetroCluster configuration, you should not initiate aggregate relocation while disaster recovery operations (switchover, healing, or switchback) are in progress. • If you have a MetroCluster configuration and initiate aggregate relocation on a switched-over aggregate, the operation might fail because it exceeds the DR partner's volume limit count. • You should not initiate aggregate relocation on aggregates that are corrupt or undergoing maintenance. • For All-Flash Optimized FAS80xx-series systems, both nodes in the HA pair must have the All- Flash Optimized personality enabled. Because the All-Flash Optimized configuration supports only SSDs, if one node in the HA pair has HDDs or array LUNs (and therefore, is not configured with the All-Flash Optimized personality), you cannot perform an aggregate relocation from that node to the node with the All- Flash Optimized personality enabled. • If the source node is used by an Infinite Volume with SnapDiff enabled, you must perform additional steps before initiating the aggregate relocation and then perform the relocation in a specific manner. You must ensure that the destination node has a namespace mirror constituent and make decisions about relocating aggregates that include namespace constituents. Clustered Data ONTAP 8.3 Infinite Volumes Management Guide • Before initiating the aggregate relocation, you should save any core dumps on the source and destination nodes. Steps 1. View the aggregates on the node to confirm which aggregates to move and ensure they are online and in good condition: storage aggregate show -node source-node Example The following command shows six aggregates on the four nodes in the cluster. All aggregates are online. Node1 and Node3 form an HA pair and Node2 and Node4 form an HA pair. cluster::> storage aggregate show Aggregate Size Available Used% State #Vols Nodes RAID Status --------- -------- --------- ----- ------- ------ ------ ----------- aggr_0 239.0GB 11.13GB 95% online 1 node1 raid_dp, Managing aggregates | 185
  • 186. normal aggr_1 239.0GB 11.13GB 95% online 1 node1 raid_dp, normal aggr_2 239.0GB 11.13GB 95% online 1 node2 raid_dp, normal aggr_3 239.0GB 11.13GB 95% online 1 node2 raid_dp, normal aggr_4 239.0GB 238.9GB 0% online 5 node3 raid_dp, normal aggr_5 239.0GB 239.0GB 0% online 4 node4 raid_dp, normal 6 entries were displayed. 2. Issue the command to start the aggregate relocation: storage aggregate relocation start -aggregate-list aggregate-1, aggregate-2... -node source-node -destination destination-node The following command moves the aggregates aggr_1 and aggr_2 from Node1 to Node3. Node3 is Node1's HA partner. The aggregates can be moved only within the HA pair. cluster::> storage aggregate relocation start -aggregate-list aggr_1, aggr_2 -node node1 -destination node3 Run the storage aggregate relocation show command to check relocation status. node1::storage aggregate> 3. Monitor the progress of the aggregate relocation with the storage aggregate relocation show command: storage aggregate relocation show -node source-node Example The following command shows the progress of the aggregates that are being moved to Node3: cluster::> storage aggregate relocation show -node node1 Source Aggregate Destination Relocation Status ------ ----------- ------------- ------------------------ node1 aggr_1 node3 In progress, module: wafl aggr_2 node3 Not attempted yet 2 entries were displayed. node1::storage aggregate> When the relocation is complete, the output of this command shows each aggregate with a relocation status of Done. Related tasks Relocating ownership of aggregates used by Infinite Volumes on page 156 186 | Physical Storage Management Guide
  • 187. Commands for aggregate relocation There are specific Data ONTAP commands for relocating aggregate ownership within an HA pair. If you want to... Use this command... Start the aggregate relocation process. storage aggregate relocation start Monitor the aggregate relocation process storage aggregate relocation show Related information Clustered Data ONTAP 8.3 Commands: Manual Page Reference Key parameters of the storage aggregate relocation start command The storage aggregate relocation start command includes several key parameters used when relocating aggregate ownership within an HA pair. Parameter Meaning -node nodename Specifies the name of the node that currently owns the aggregate -destination nodename Specifies the destination node where aggregates are to be relocated -aggregate-list aggregate name Specifies the list of aggregate names to be relocated from source node to destination node (This parameter accepts wildcards) -override-vetoes true|false Specifies whether to override any veto checks during the relocation operation Use of this option can potentially lead to longer client outage, or aggregates and volumes not coming online after the operation. Managing aggregates | 187
  • 188. Parameter Meaning -relocate-to-higher-version true| false Specifies whether the aggregates are to be relocated to a node that is running a higher version of Data ONTAP than the source node • You cannot perform an aggregate relocation from a node running Data ONTAP 8.2 to a node running Data ONTAP 8.3 or higher using the ‑relocate‑to‑higher-version true parameter. You must first upgrade the source node to Data ONTAP 8.2.1 or higher before you can perform an aggregate relocation operation using this parameter. Similarly, you must upgrade to Data ONTAP 8.2.1 or higher before you can upgrade to Data ONTAP 8.3. • Although you can perform aggregate relocation between nodes running different minor versions of Data ONTAP (for example, 8.2.1 to 8.2.2, or 8.2.2 to 8.2.1), you cannot perform an aggregate relocation operation from a higher major version to a lower major version (for example, 8.3 to 8.2.2). -override-destination-checks true| false Specifies if the aggregate relocation operation should override the check performed on the destination node Use of this option can potentially lead to longer client outage, or aggregates and volumes not coming online after the operation. Veto and destination checks during aggregate relocation In aggregate relocation operations, Data ONTAP determines whether aggregate relocation can be completed safely. If aggregate relocation is vetoed, you must check the EMS messages to determine the cause. Depending on the reason or reasons, you can decide whether you can safely override the vetoes. The storage aggregate relocation show command displays the aggregate relocation progress and shows which subsystem, if any, vetoed the relocation. Soft vetoes can be overridden, but hard vetoes cannot be, even if forced. You can review the EMS details for any giveback vetoes by using the following command: 188 | Physical Storage Management Guide
  • 189. event log show -node * -event gb* You can review the EMS details for aggregate relocation by using the following command: event log show -node * -event arl* The following tables summarize the soft and hard vetoes, along with recommended workarounds: Veto checks during aggregate relocation Vetoing subsystem module Workaround Vol Move Relocation of an aggregate is vetoed if any volumes hosted by the aggregate are participating in a volume move that has entered the cutover state. Wait for the volume move to complete. If this veto is overridden, cutover resumes automatically once the aggregate relocation completes. If aggregate relocation causes the move operation to exceed the number of retries (the default is 3), then the user needs to manually initiate cutover using the volume move trigger-cutover command. Backup Relocation of an aggregate is vetoed if a dump or restore job is in progress on a volume hosted by the aggregate. Wait until the dump or restore operation in progress is complete. If this veto is overridden, the backup or restore operation is aborted and must be restarted by the backup application. Lock manager To resolve the issue, gracefully shut down the CIFS applications that have open files, or move those volumes to a different aggregate. Overriding this veto results in loss of CIFS lock state, causing disruption and data loss. Lock Manager NDO Wait until the locks are mirrored. This veto cannot be overridden; doing so disrupts Microsoft Hyper-V virtual machines. RAID Check the EMS messages to determine the cause of the veto: If disk add or disk ownership reassignment operations are in progress, wait until they complete. If the veto is due to a mirror resync, a mirror verify, or offline disks, the veto can be overridden and the operation restarts after giveback. Managing aggregates | 189
  • 190. Destination checks during aggregate relocation Vetoing subsystem module Workaround Disk Inventory Relocation of an aggregate fails if the destination node is unable to see one or more disks belonging to the aggregate. Check storage for loose cables and verify that the destination can access disks belonging to the aggregate being relocated. This check cannot be overridden. WAFL Relocation of an aggregate fails if the relocation would cause the destination to exceed its limits for maximum volume count or maximum volume size. This check cannot be overridden. Lock Manager NDO Relocation of an aggregate fails if: • The destination does not have sufficient lock manager resources to reconstruct locks for the relocating aggregate. • The destination node is reconstructing locks. Retry aggregate relocation after a few minutes. This check cannot be overridden. Lock Manager Permanent relocation of an aggregate fails if the destination does not have sufficient lock manager resources to reconstruct locks for the relocating aggregate. Retry aggregate relocation after a few minutes. This check cannot be overridden. RAID Check the EMS messages to determine the cause of the failure: • If the failure is due to an aggregate name or UUID conflict, troubleshoot and resolve the issue. This check cannot be overridden. Relocating an aggregate fails if the relocation would cause the destination to exceed its limits for maximum aggregate count, system capacity, or aggregate capacity. You should avoid overriding this check. Assigning aggregates to SVMs If you assign one or more aggregates to a Storage Virtual Machine (SVM, formerly known as Vserver), then you can use only those aggregates to contain volumes for that SVM. Assigning 190 | Physical Storage Management Guide
  • 191. aggregates to your SVMs is particularly important in a multi-tenancy environment or when you use Infinite Volumes. Before you begin The SVM and the aggregates you want to assign to that SVM must already exist. About this task Assigning aggregates to your SVMs helps you keep your SVMs isolated from each other; this is especially important in a multi-tenancy environment. If you use Infinite Volumes, or plan to use them in the future, you must assign aggregates to your SVMs to keep your Infinite Volumes from impacting each other and any FlexVol volumes in your cluster. Steps 1. Check the list of aggregates already assigned to the SVM: vserver show -fields aggr-list The aggregates currently assigned to the SVM are displayed. If there are no aggregates assigned, “-” is displayed. 2. Add or remove assigned aggregates, depending on your requirements: If you want to... Use this command... Assign additional aggregates vserver add-aggregates Unassign aggregates vserver remove-aggregates The listed aggregates are assigned to or removed from the SVM. If the SVM already has volumes that use an aggregate that is not assigned to the SVM, a warning message is displayed, but the command is completed successfully. Any aggregates that were already assigned to the SVM and that were not named in the command are unaffected. Example In the following example, the aggregates aggr1 and aggr2 are assigned to SVM svm1: vserver add-aggregates -vserver svm1 -aggregates aggr1,aggr2 Methods to create space in an aggregate If an aggregate runs out of free space, various problems can result that range from loss of data to disabling a volume's guarantee. There are multiple ways to make more space in an aggregate. All of the methods have various consequences. Prior to taking any action, you should read the relevant section in the documentation. Managing aggregates | 191
  • 192. The following are some common ways to make space in an aggregate, in order of least to most consequences: • Add disks to the aggregate. • Move some volumes to another aggregate with available space. • Shrink the size of volume-guaranteed volumes in the aggregate. You can do this manually or with the autoshrink option of the autosize capability. • Change volume guarantee types to none on volumes that are using large amounts of space (large volume-guaranteed volumes with large reserved files) so that the volumes take up less space in the aggregate. A volume with a guarantee type of none has a smaller footprint in the aggregate than a volume with a guarantee type of volume. The Volume Guarantee row of the volume show- footprint command output shows whether a volume is reserving a large amount of space in the aggregate due to its guarantee. • Delete unneeded volume Snapshot copies if the volume's guarantee type is none. • Delete unneeded volumes. • Enable space-saving features, such as deduplication or compression. • (Temporarily) disable features that are using a large amount of metadata (visible with the volume show-footprint command). Determining which volumes reside on an aggregate You might need to determine which FlexVol volumes or Infinite Volume constituents reside on an aggregate before performing operations on the aggregate, such as relocating it or taking it offline. About this task Infinite Volume constituents are somewhat similar to FlexVol volumes, but you usually do not manage them directly. For more information about Infinite Volumes and constituents, see the Clustered Data ONTAP Infinite Volumes Management Guide. Step 1. Enter the appropriate command, depending on whether your system has Infinite Volumes: If your system... Then use this command... Does not have Infinite Volumes volume show -aggregate aggregate_name 192 | Physical Storage Management Guide
  • 193. If your system... Then use this command... Has Infinite Volumes volume show -is-constituent * -aggregate aggregate_name All volumes (and, if you have Infinite Volumes, constituents) that reside on the specified aggregate are displayed. Determining whether a Flash Pool aggregate is using an SSD storage pool You manage Flash Pool aggregates differently when they use SSD storage pools to provide their cache than when they use discrete SSDs. Step 1. Display the aggregate's drives by RAID group: storage aggregate show-status aggr_name If the aggregate is using one or more SSD storage pools, the value for the Position column for the SSD RAID groups is displayed as Shared, and the name of the storage pool is displayed next to the RAID group name. Commands for managing aggregates You use the storage aggregate command to manage your aggregates. If you want to... Use this command... Display the size of the cache for all Flash Pool aggregates storage aggregate show -fields hybrid-cache-size-total -hybrid- cache-size-total >0 Display disk information and status for an aggregate storage aggregate show-status Display spare disks by node storage aggregate show-spare-disks Display the root aggregates in the cluster storage aggregate show -has-mroot true Display basic information and status for aggregates storage aggregate show Bring an aggregate online storage aggregate online Managing aggregates | 193
  • 194. If you want to... Use this command... Delete an aggregate storage aggregate delete Put an aggregate into the restricted state storage aggregate restrict Rename an aggregate storage aggregate rename Take an aggregate offline storage aggregate offline Change the RAID type for an aggregate storage aggregate modify -raidtype Related information Clustered Data ONTAP 8.3 Commands: Manual Page Reference 194 | Physical Storage Management Guide
  • 195. Storage limits There are limits for storage objects that you should consider when planning and managing your storage architecture. Limits are listed in the following sections: • Aggregate limits • RAID group limits Aggregate limits Storage object Limit Native storage Storage arrays Virtual storage (Data ONTAP-v) Aggregates Maximum per node1 100 100 60 Maximum size2 Model- dependent Model- dependent 16 TB Minimum size3 RAID-DP: 5 disks RAID4: 3 disks Model- dependent 1 disk Aggregates (mirrored) Maximum suggested per node4 64 64 N/A RAID groups Maximum per aggregate 150 150 60 Notes: 1. In an HA configuration, this limit applies to each node individually, so the overall limit for the pair is doubled. 2. See the Hardware Universe. 3. For root aggregates using physical disks, the minimum size is 3 disks for RAID-DP and 2 disks for RAID4. See the Hardware Universe for the minimum aggregate size for storage arrays. 195
  • 196. 4. You can create more than 64 mirrored aggregates on a node, but doing so could cause plex synchronization problems after certain types of failures. RAID group limits For maximum and default RAID group sizes, see the Hardware Universe. Limit Native storage Storage arrays Virtual storage (Data ONTAP-v) Maximum per node 400 400 60 Maximum per aggregate 150 150 60 196 | Physical Storage Management Guide
  • 197. Copyright information Copyright © 1994–2015 NetApp, Inc. All rights reserved. Printed in the U.S. No part of this document covered by copyright may be reproduced in any form or by any means— graphic, electronic, or mechanical, including photocopying, recording, taping, or storage in an electronic retrieval system—without prior written permission of the copyright owner. Software derived from copyrighted NetApp material is subject to the following license and disclaimer: THIS SOFTWARE IS PROVIDED BY NETAPP "AS IS" AND WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, WHICH ARE HEREBY DISCLAIMED. IN NO EVENT SHALL NETAPP BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. NetApp reserves the right to change any products described herein at any time, and without notice. NetApp assumes no responsibility or liability arising from the use of products described herein, except as expressly agreed to in writing by NetApp. The use or purchase of this product does not convey a license under any patent rights, trademark rights, or any other intellectual property rights of NetApp. The product described in this manual may be protected by one or more U.S. patents, foreign patents, or pending applications. RESTRICTED RIGHTS LEGEND: Use, duplication, or disclosure by the government is subject to restrictions as set forth in subparagraph (c)(1)(ii) of the Rights in Technical Data and Computer Software clause at DFARS 252.277-7103 (October 1988) and FAR 52-227-19 (June 1987). 197
  • 198. Trademark information NetApp, the NetApp logo, Go Further, Faster, ASUP, AutoSupport, Campaign Express, Cloud ONTAP, clustered Data ONTAP, Customer Fitness, Data ONTAP, DataMotion, Fitness, Flash Accel, Flash Cache, Flash Pool, FlashRay, FlexArray, FlexCache, FlexClone, FlexPod, FlexScale, FlexShare, FlexVol, FPolicy, GetSuccessful, LockVault, Manage ONTAP, Mars, MetroCluster, MultiStore, NetApp Insight, OnCommand, ONTAP, ONTAPI, RAID DP, SANtricity, SecureShare, Simplicity, Simulate ONTAP, Snap Creator, SnapCopy, SnapDrive, SnapIntegrator, SnapLock, SnapManager, SnapMirror, SnapMover, SnapProtect, SnapRestore, Snapshot, SnapValidator, SnapVault, StorageGRID, Tech OnTap, Unbound Cloud, and WAFL are trademarks or registered trademarks of NetApp, Inc., in the United States, and/or other countries. A current list of NetApp trademarks is available on the web at http://www.netapp.com/us/legal/netapptmlist.aspx. Cisco and the Cisco logo are trademarks of Cisco in the U.S. and other countries. All other brands or products are trademarks or registered trademarks of their respective holders and should be treated as such. 198 | Physical Storage Management Guide
  • 199. How to send comments about documentation and receive update notification You can help us to improve the quality of our documentation by sending us your feedback. You can receive automatic notification when production-level (GA/FCS) documentation is initially released or important changes are made to existing production-level documents. If you have suggestions for improving this document, send us your comments by email to doccomments@netapp.com. To help us direct your comments to the correct division, include in the subject line the product name, version, and operating system. If you want to be notified automatically when production-level documentation is released or important changes are made to existing production-level documents, follow Twitter account @NetAppDoc. You can also contact us in the following ways: • NetApp, Inc., 495 East Java Drive, Sunnyvale, CA 94089 U.S. • Telephone: +1 (408) 822-6000 • Fax: +1 (408) 822-4501 • Support telephone: +1 (888) 463-8277 199
  • 200. Index A ACP how you use to increase storage availability for SAS- connected disk shelves 24 active-active configurations setting up with root-data partitioning 39 adding key management servers 90 aggregate ownership relocation of 182 aggregate relocation benefits of 183 commands for 187 effect on root-data partitioning 184 effect on shared disks 184 how it works 183 Infinite Volumes 156 monitoring progress of 188 overriding a veto of 188 aggregate show-space command how to determine aggregate space usage by using 148 aggregates adding physical drives or array LUNs to 163 assigning to SVMs 190 associated with Infinite Volumes 154 changing size of RAID groups for 115 commands for displaying space usage information 53 commands for managing 193 configuration requirements for multi-disk carrier shelves 29 considerations for sizing RAID groups 114 considerations for using disks from multi-disk carriers in 29 creating Flash Pool 171 creating Flash Pool using SSD storage pools 173 creating SSD storage pool for Flash Pool 140 creating using physical drives 159 creating using root-data partitioning 161 creating using shared HDDs 161 description and characteristics of 127 determination of checksum type of array LUN 148 determining candidacy and optimal cache size for Flash Pool 175 determining drive information for 181 determining RAID group information for 181 determining which volumes reside on 192 disk speeds supported by Data ONTAP 13 effect of SVM on selection 144 Flash Pool, defined 130 Flash Pool, determining volume write-caching eligibility for 178 Flash Pool, determining whether using a storage pool 193 Flash Pool, how they work 130 how associated with SVMs with Infinite Volume 154 how cache capacity is calculated for Flash Pool 133 how drive checksum types affect management 14 how root-data partitioning affects storage management 31 how storage classes use for Infinite Volumes 153 how storage pools increase cache allocation flexibility for Flash Pool 137 how to determine space usage in 148 how you use storage pools with Flash Pool 139 increasing size of, when they use shared HDDs 165 increasing the size of with physical drives 163 introduction to managing 159 maximum and minimum size of 195 maximum per node 195 maximum size, methods of calculating 13 methods of creating space in 191 mirrored, explained 129 mirrored, maximum per node 195 ownership change 184 relocating ownership of 156 relocation of 183, 184 replacing disks used in 43 requirements and best practices for using storage pools with Flash Pool 139 requirements for Infinite Volumes 152 requirements for storage classes 153 requirements for using Flash Pool 131 root, which drives are partitioned for use in 32 rules about mixing storage types in 147 rules for mixing HDD types in 146 rules for storage array families 147 sharing between FlexVol volumes and Infinite Volumes 152 tips for creating and backing up, for sensitive data 21 200 | Physical Storage Management Guide
  • 201. unmirrored, explained 127 ways to use disks with mixed speeds in 145 what happens when adding storage to 170 All-Flash Optimized personality how it affects node behavior 25 Alternate Control Path See ACP array LUN ownership how it works 66 array LUNs adding to aggregates 163 considerations for sizing RAID groups for 114 how to control selection from heterogeneous storage 145 reasons to assign ownership of 54, 66 systems running Data ONTAP that can use 64 verifying back-end configuration for 72 assigning aggregates 154 ATA drives how Data ONTAP reports disk types 10 authentication keys changing 96 deleting 98 how Storage Encryption uses 82 removing management server that stores 93 retrieving 97 autoassignment See automatic ownership assignment Automated Workflow Analyzer See AWA automatic ownership assignment configuring for disks 61 guidelines for disks 58 how it works for disks 56 when it is invoked 57 which policy to use for disk 56 automatic RAID-level scrubs changing the schedule for 122 availability how Data ONTAP uses RAID to ensure data 108 AWA determining Flash Pool optimal cache size by using 175 AZCS type checksums configuration rules 14 effect on aggregate management 14 effect on spare management 14 B back-end configurations verifying 72 BCS type checksums configuration rules 14 effect on aggregate management 14 effect on spare management 14 benefits of Storage Encryption 83 best practices for using storage pools 139 block checksum type changing for array LUNs 76 BSAS drives how Data ONTAP reports disk types 10 C cache capacity how calculated for Flash Pool aggregates 133 cache size determining impact to, when adding SSDs to storage pool 143 determining optimal for Flash Pool aggregates 175 cache storage requirements and best practices for using storage pools for Flash Pool aggregate 139 caches comparison of Flash Pool and Flash Cache 133 caching policies modifying 136 using Flash Pool aggregates 135 capacity how it is allocated in new Infinite Volumes 155 methods of calculating aggregate and system 13 carriers determining when to remove multi-disk 28 how Data ONTAP avoids RAID impact when removing multi-disk 27 spare requirements for multi-disk 28 certificates installing replacement SSL 99 installing SSL, on storage systems 87 preventing SSL expiration issues 98 removing old SSL 99 SSL requirements 87 changing authentication keys 96 RAID group size 115 Index | 201
  • 202. RAID type for Flash Pool cache 180 checksums checking the type 76 configuration rules 14 rules for aggregates 148 type, changing for array LUNs 76 type, effect on aggregate and spare management 14 commands aggregate management, list of 193 disk management, list of 51 for displaying aggregate space usage information 53 for displaying FlexVol volume space usage information 53 for displaying information about storage shelves 52 how you use wildcard character with disk ownership 62 SSD storage pool management, list of 143 storage aggregate 51, 53, 193 storage disk 51 storage pool 143 storage pool management, list of 143 storage shelves, displaying information about 52 volume show-footprint 53 volume show-space 53 volume snapshot 53 comments how to send feedback about documentation 199 composition changing array LUN 78 configuring automatic ownership assignment of disks 61 connection types how disks can be combined for SAS 12 supported storage 12 constituents determining which ones reside on an aggregate 192 continuous media scrubbing how Data ONTAP uses, to prevent media errors 24 impact on system performance 24 reasons it should not replace scheduled RAID-level disk scrubs 24 core files spare disk requirement for 117 creating aggregates using physical drives 159 Flash Pool aggregates 171 Flash Pool aggregates using SSD storage pools 173 current owner disk ownership type, defined 54 D data how Data ONTAP uses RAID to protect and ensure availability 108 tips for creating and backing up aggregates containing sensitive 21 using sanitization to remove from disks 48 data at rest introduction to securing with Storage Encryption 81 data disks removing 47 data integrity how RAID-level disk scrubs verify 122 Data ONTAP disk types comparison with industry standard 10 Data ONTAP Edge how disk ownership works 57 Data ONTAP-v how disk ownership works 57 data protection in case of disk loss or theft 83 through emergency shredding 84 when moving disks to end-of-life 84 when returning disks to vendors 83 data reconstruction controlling performance impact of RAID 124 data shredding performing emergency, on disks using Storage Encryption 104 degraded mode 119 deleting authentication keys 98 destroying data on disks using Storage Encryption 102 disk performance monitors 21 disk connection types how disks can be combined for SAS 12 disk operations with SEDs 82 disk ownership application to array LUNs 55, 66, 68 application to disks 55, 66 assigning array LUNs 71 configuring automatic assignment 61 how it works 66 how it works for Data ONTAP Edge 57 how it works for Data ONTAP-v 57 ownership 202 | Physical Storage Management Guide
  • 203. removing array LUN ownership 79 removing array LUN ownership 79 disk ownership commands how you use wildcard character with 62 Disk Qualification Package when you need to update 42 disk remove -w removing an array LUN 79 disk sanitization introduction to how it works 19 process described 19 when it cannot be performed 20 disk scrubbing reasons continuous media scrubbing should not replace scheduled RAID-level 24 disk shelves aggregate configuration requirements for multi-disk carrier 29 commands for displaying information about 52 configuration requirements for multi-disk carrier 29 how you use ACP to increase storage availability for SAS-connected 24 requirements for using multi-disk carrier 27 disk slicing understanding for entry-level platforms 30 disk types how to control selection from heterogeneous storage 145 disks adding to a node 41 adding to aggregates 163 assigning ownership for 58 commands for managing 51 considerations for removing from storage systems 45 considerations for using, from multi-disk carriers in aggregates 29 data, converting to spare 44 determining information about, for aggregates 181 displaying information about Storage Encryption 95 evacuation process, about 27 guidelines for assigning ownership 58 how automatic ownership assignment works 56 how available for Data ONTAP use 55, 66 how Data ONTAP handles failed, with available hot spares 120 how Data ONTAP handles failed, with no available hot spare 121 how Data ONTAP reduces failures using Rapid RAID Recovery 21 how Data ONTAP reports types 10 how low spare warnings can help you manage spare 120 how RAID-level scrubs verify data integrity 122 how root-data partitioning affects storage management 31 how shared HDDs work 30 how they can be combined for SAS connection type 12 how to control selection from heterogeneous storage 145 introduction to how DATA ONTAP works with heterogeneous storage 145 introduction to managing ownership for 54 loop IDs for FC-AL connected, about 18 managing using Data ONTAP 10 matching spares defined 118 minimum required hot spare 117 performing emergency data shredding on Storage Encryption 104 physical, adding to aggregates 163 RAID drive types, defined 18, 113 RAID protection levels for 108 reasons to assign ownership of 54, 66 removing data 47 removing failed 45 removing hot spares 46 removing ownership from 60 replacing in aggregate 43 replacing self-encrypting 44 requirements for using partitioned 34 rules for mixing HDD types in aggregates 146 sanitization process described 19 sanitization, what happens if interrupted 20 sanitizing 102 setting state to end-of-life 103 slicing, understanding for entry level platforms 30 spare requirements for multi-disk carrier 28 spare, appropriate 118 speeds supported by Data ONTAP 13 SSD and HDD capability differences 26 stopping sanitization 51 types of ownership for 54 using sanitization to remove data from 48 ways to mix speed of, in aggregates 145 what happens when Data ONTAP takes them offline 21 when automatic ownership assignment is invoked 57 when sanitization cannot be performed 20 when they can be put into maintenance center 23 Index | 203
  • 204. when you need to update the Disk Qualification Package for 42 which autoassignment policy to use for 56 displaying key management server information 92 key management server status 91 Storage Encryption disk information 95 documentation how to receive automatic notification of changes to 199 how to send feedback about 199 DQP See Disk Qualification Package DR home owner disk ownership type, defined 54 drive errors how the maintenance center helps prevent 22 drive types RAID, defined 18, 113 drives considerations for sizing RAID groups for 114 how Data ONTAP handles failed, with available hot spares 120 how Data ONTAP handles failed, with no available hot spares 121 how low spare warnings can help you manage spare 120 name formats 14 pre-cluster name formats 15 rules for mixing types in Flash Pool aggregates 147 which are partitioned for root-data partitioning 32 See also disks E emergency data shredding data protection through 84 performing on disks using Storage Encryption 104 end-of-life setting disk state to 103 entry-level platforms understanding root-data partitioning for 30 errors how Data ONTAP uses media scrubbing to prevent media 24 how the maintenance center helps prevent drive 22 evacuation process for disks, about 27 existing storage pools considerations for adding SSDs to 142 external key management servers defined 81 displaying information about 92 F failed disks removing 45 family defined 147 FC storage connection type how disks can be combined for 12 support for 12 FCAL drives how Data ONTAP reports disk types 10 feedback how to send comments about documentation 199 Fibre Channel See FC Flash Cache compared with Flash Pool aggregates 133 Flash Pool aggregates caching policies 135 changing RAID type 180 compared with Flash Cache 133 creating 171 creating SSD storage pool for 140 creating using SSD storage pools 173 defined 130 determining candidacy and optimal cache size for 175 determining whether using a storage pool 193 how cache capacity is calculated for 133 how storage pools increase cache allocation flexibility for 137 how they work 130 how you use storage pools with 139 RAID type, considerations for 132 requirements and best practices for using storage pools with 139 requirements for using 131 rules for mixing drive types in 147 spare management, considerations for 132 volume write-caching eligibility, determining 178 Flash Pool SSD partitioning how it increases cache allocation flexibility for Flash Pool aggregates 137 FlexVol volumes aggregate sharing with Infinite Volumes 152 204 | Physical Storage Management Guide
  • 205. commands for displaying space usage information 53 creating aggregates using physical drives 159 determining which ones reside on an aggregate 192 effect of SVM on aggregate selection 144 formats drive name 14 pre-cluster drive name 15 FSAS drives how Data ONTAP reports disk types 10 G groups RAID, how they work 113 guidelines assigning disk ownership 58 H hard disk drives See HDDs HDD RAID groups sizing considerations for 114 HDDs assigning ownership for shared 59 capability differences with SSDs 26 creating aggregates using shared 161 increasing size of aggregates that use shared 165 rules for mixing types in aggregates 146 shared, how they work 30 speeds supported by Data ONTAP 13 standard layouts for shared 32 heterogeneous storage how to control disk selection from 145 introduction to how Data ONTAP works with 145 high-performance aggregates Flash Pool, defined 130 home owner disk ownership type, defined 54 hot spares appropriate 118 defined 117 how Data ONTAP handles failed disks with available 120 how Data ONTAP handles failed disks with no available 121 matching, defined 118 minimum needed 117 removing 46 what disks can be used as 118 hybrid aggregates See Flash Pool aggregates I IDs about loop, for FC-AL connected disks 18 increasing aggregate size using physical drives 163 Infinite Volumes aggregate relocation 156 aggregate requirements 152 associated aggregates 154 capacity allocation 155 creating aggregates using physical drives 159 determining which constituents reside on an aggregate 192 how storage classes use aggregates for 153 how to determine space usage for constituents 150 relocating aggregates 156 space allocation 155 InfiniteVol See Infinite Volumes information how to send feedback about improving documentation 199 initialization performing on node to configure root-data partitioning 34 installing replacement SSL certificates 99 SSL certificates on storage systems 87 K Key Management Interoperability Protocol using for communication with key management servers 81 key management servers adding 90 displaying information about 92 displaying status 91 external, defined 81 introduction to using SSL for secure communication 86 precautions taken against unreachability during boot process 94 removing 93 verifying links 91 Index | 205
  • 206. keys changing authentication 96 how Storage Encryption uses authentication 82 retrieving authentication 97 KMIP See Key Management Interoperability Protocol L layouts standard shared HDD 32 levels RAID protection, for disks 108 licenses for using array LUNs 65 installing for array LUN use 65 V_StorageAttach 65 limitations Storage Encryption 84 limits aggregate storage 195 FlexClone file and LUN storage 195 RAID group storage and size 195 volume storage 195 loop IDs about FC-AL connected disk 18 loops configuring automatic ownership assignment for 61 low spare warnings how they can help you manage spare drives 120 LUNs (array) assigning ownership of 71 changing checksum type 76 changing ownership assignment 73 changing size or composition 78 checking the checksum type of 76 Data ONTAP owning 68 Data ONTAP RAID groups with 116 examples of when Data ONTAP can use 69 how available for Data ONTAP use 55, 66 managing through Data ONTAP 64 names format of 74 prerequisites to changing composition 77 prerequisites to changing size 77 RAID protection for 109 reasons to assign ownership of 54, 66 reasons you might assign to a system 68 requirements before removing a system running Data ONTAP from service 79 rules about mixing storage types in aggregates 147 setting them up in Data ONTAP 64 systems running Data ONTAP that can use 64 verifying back-end configuration for 72 LUNs, array See LUNs (array) M maintenance center how it helps prevent drive errors 22 when disks go into 23 management servers adding key 90 displaying information about key 92 removing authentication key 93 verifying server links of key 91 managing Storage Encryption 90 manual RAID-level scrubs how to run 123 matching spare disks defined 118 media errors how Data ONTAP uses media scrubbing to prevent 24 media scrubbing how Data ONTAP uses, to prevent media errors 24 impact on system performance 24 reasons it should not replace scheduled RAID-level disk scrubs 24 mirror verification controlling performance impact 126 mirrored aggregates explained 129 modifying caching policies 136 MSATA drives how Data ONTAP reports disk types 10 MSIDs rekeying SEDs to 100 multi-disk carrier shelves aggregate configuration requirements for 29 configuration requirements for 29 in aggregates, considerations for using disks from 29 requirements for using 27 multi-disk carriers determining when to remove 28 how Data ONTAP handles when removing 27 spare requirements for 28 206 | Physical Storage Management Guide
  • 207. N name format array LUNs 74 names formats for drive 14 formats for pre-cluster drive 15 new storage pools considerations for adding SSDs to 142 NL-SAS drives how Data ONTAP reports disk types 10 node behavior how All-Flash Optimized personality affects 25 nodes adding disks to 41 how to control disk selection from heterogeneous storage on 145 initializing for root-data partitioning 34 O offline what happens when Data ONTAP takes disks 21 original owner disk ownership type, defined 54 owner disk ownership type, defined 54 ownership assigning array LUNs 71 assigning for disks 58 assigning for shared HDDs 59 automatically assigning to a stack or shelf 61 disk, types of 54 guidelines for assigning disk 58 how it works for disks and array LUNs 66 introduction to managing, for disks 54 reasons to assign disk and array LUN 54, 66 removing from disks 60 ownership assignment when it is invoked for disks 57 ownership commands how you use wildcard character with disk 62 P partitioning how Flash Pool SSD increases cache allocation flexibility for 137 root-data, assigning ownership for HDDs partitioned for 59 root-data, creating aggregates using 161 root-data, how it works 30 root-data, how storage management is affected by 31 root-data, increasing the size of an aggregate using 165 root-data, requirements for using 34 root-data, standard layouts for 32 root-data, understanding 30 root-data, which drives are partitioned for 32 setting up active-active configuration on node using root-data 39 partitions correcting misaligned spare 169 performance controlling impact of RAID data reconstruction 124 impact of media scrubbing on system 24 persistent reservations releasing all 79 physical drives using to create aggregates 159 physical secure IDs See PSIDs platforms understanding root-data partitioning for entry-level 30 plex resynchronization controlling performance impact of 125 plexes mirrored aggregate, explained 129 policies autoassignment, which to use for disks 56 pools considerations for when to use SSD storage 138 requirements and best practices for using storage 139 protection how Data ONTAP uses to RAID for data 108 RAID levels for disks 108 PSIDs how the factory resets SEDs with 105 SEDs that include the functionality 106 using to reset SED to factory original settings 106 R RAID avoiding impact to, when replacing multi-disk carriers 27 data reconstruction, controlling performance impact 124 drive types defined 18, 113 Index | 207
  • 208. how Data ONTAP to protect data and data availability 108 how disk scrubs verify data integrity 122 operations, controlling performance impact 123 protection levels for disks 108 protection with SyncMirror and 110 scrub, controlling performance impact 124 scrubs, changing the schedule for 122 type, changing for Flash Pool cache 180 type, determining for Flash Pool cache 132 RAID groups changing size of 115 definition 113 determining information about, for aggregates 181 how they work 113 maximum per aggregate 195 maximum per node 196 naming convention 114 sizing considerations for 114 what happens when adding storage to aggregates in 170 with array LUNs, considerations 116 RAID protection for array LUNs 109 RAID-DP described 108 RAID-level scrubs automatic schedule, changing 122 how to run manual 123 reasons media scrubbing should not replace scheduled 24 raid.timeout option considerations for changing 121 RAID0 how Data ONTAP uses for array LUNs 109 use by Data ONTAP 109 RAID4 described 109 Rapid RAID Recovery how Data ONTAP reduces disk failures using 21 rekeying SEDs to MSID 100 relocating aggregates Infinite Volumes 156 relocation aggregate ownership 182, 184 of aggregates 182–184 removing data disks 47 data, using disk sanitization 48 failed disks 45 hot spare disks 46 key management servers 93 multi-disk carriers, determining when it is safe 28 old SSL certificates 99 replacing disks in aggregates 43 requirements Flash Pool aggregate use 131 for using root-data partitioning 34 for using storage pools 139 Infinite Volumes, aggregate 152 resetting SEDs to factory original settings 106 resynchronization controlling performance impact of plex 125 retrieving authentication keys 97 root-data partitioning assigning ownership for HDDs using 59 correcting misaligned spare partitions for 169 creating aggregates using 161 effect on aggregate relocation 184 how it works 30 how storage management is affected by 31 increasing the size of an aggregate using 165 initializing node to configure 34 requirements for using 34 setting up active-active configuration on node using 39 standard layouts for 32 understanding 30 which drives are partitioned for 32 rules for mixing drive types in Flash Pool aggregates 147 for mixing HDD types in aggregates 146 S sanitization disk process described 19 disk, introduction to how it works 19 stopping disk 51 tips for creating and backing up aggregates containing sensitive data 21 using to remove data from disks 48 what happens if interrupted 20 when it cannot be performed 20 sanitizing disks 102 208 | Physical Storage Management Guide
  • 209. SAS storage connection type, support for 12 SAS drives how Data ONTAP reports disk types 10 SAS-connected shelves how disks can be combined for 12 how you use ACP to increase storage availability for 24 SATA drives how Data ONTAP reports disk types 10 scrubbing how Data ONTAP uses media, to prevent media errors 24 impact of media, on system performance 24 media, reasons it should not replace scheduled RAID-level disk scrubs 24 scrubs changing automatic schedule for 122 controlling performance impact of RAID 124 how to run manual RAID-level 123 RAID-level, how they verify data integrity 122 secure communication introduction to using SSL for 86 SEDs disk operations with 82 how Storage Encryption works with 82 how the factory resets with PSID 105 replacing 44 resetting to factory original settings 106 returning to unprotected mode 100 that have PSID functionality 106 self-encrypting disks See SEDs serial-attached SCSI See SAS servers adding key management 90 displaying information about key management 92 removing authentication key management 93 verifying server links of key management 91 setting up array LUNs 64 Storage Encryption 85 setup wizards running the Storage Encryption 88 shared HDDs assigning ownership for 59 creating aggregates using 161 how they work 30 increasing size of aggregates that use 165 understanding 30 shared layouts standard HDD 32 shared SSDs See storage pools shelves aggregate configuration requirements for multi-disk carrier 29 commands for displaying information about 52 configuration requirements for multi-disk carrier 29 configuring automatic ownership assignment for 61 how you use ACP to increase storage availability for SAS-connected 24 requirements for using multi-disk carrier 27 shredding performing emergency data, on disks using Storage Encryption 104 size changing array LUN size 77 sizes changing array LUN 78 sizing RAID groups, considerations for 114 Snapshot reserve commands for displaying size of 53 solid-state disks See SSDs space commands for displaying usage information 53 how it is allocated in new Infinite Volumes 155 methods of creating in an aggregate 191 space usage how to determine and control volume, in aggregates 150 how to determine in an aggregate 148 spare array LUNs changing array LUN assignment 73 changing ownership assignment 73 checking the type 76 disk ownership 73 spare disks appropriate 118 defined 117 how checksum types affect management 14 how Data ONTAP handles failed disks with available 120 how Data ONTAP handles failed disks with no available 121 how low spare warnings can help you manage 120 managing for Flash Pool cache 132 Index | 209
  • 210. matching, defined 118 minimum needed 117 removing 46 removing ownership from 60 requirements for multi-disk carriers 28 what disks can be used as 118 spare partitions correcting misaligned 169 speeds disk, supported by Data ONTAP 13 ways to mix disk, in aggregates 145 SSD storage pools commands for managing 143 creating Flash Pool aggregates using 173 See also storage pools SSDs adding to storage pools 141 capability differences with HDDs 26 changing size of RAID groups for 115 considerations for adding to existing storage pool versus new one 142 considerations for when to use storage pools 138 determining impact to cache size of adding to storage pool 143 how Data ONTAP manages wear life 26 how Data ONTAP reports disk types 10 how used in Flash Pool aggregates 130 introduction to using 25 shared See storage pools sizing considerations for RAID groups 114 storage pools, creating 140 storage pools, determining when used by a Flash Pool aggregate 193 SSL certificates installing on storage systems 87 installing replacement 99 preventing expiration issues 98 removing old 99 requirements 87 SSL connections introduction to using for secure key management communication 86 stacks configuring automatic ownership assignment for 61 standard layouts shared HDD 32 state of disks setting to end-of-life 103 stopping disk sanitization 51 storage adding to SSD storage pools 141 how root-data partitioning affects management of 31 how to control disk selection from heterogeneous 145 what happens when adding to an aggregate 170 storage aggregate commands for displaying space information 53 for managing aggregates 193 for managing disks 51 storage aggregate relocation start command key parameters of 187 storage arrays rules about mixing in aggregates 147 storage connection types supported 12 storage disk commands for managing disks 51 Storage Encryption benefits 83 destroying data using 102 displaying disk information 95 explained 81 how it works 82 information to collect before configuring 85 installing replacement SSL certificates 99 installing SSL certificates for 87 introduction to securing data at rest with 81 introduction to using SSL for secure key management communication 86 limitations 84 managing 90 overview 81 performing emergency data shredding 104 preventing SSL certificate expiration issues 98 purpose of external key management server 81 removing old SSL certificates 99 replacing self-encrypting disks 44 running the setup wizard 88 sanitizing disks using 102 setting disk state to end-of-life 103 setting up 85 SSL certificates requirements 87 storage limits aggregate 195 FlexClone file and LUN 195 RAID group 195 volume 195 storage performance 210 | Physical Storage Management Guide
  • 211. introduction to using SSDs to increase 25 performance introduction to using SSDs to increase storage 25 storage pools adding SSDs to 141 advantages of SSD 138 commands for managing 143 considerations for adding SSDs to new versus existing 142 considerations for when to use SSD 138 creating 140 creating Flash Pool aggregates using SSD 173 determining impact to cache size of adding SSDs to 143 determining when used by a Flash Pool aggregate 193 disadvantages of SSD 138 how they increase cache allocation for Flash Pool aggregates 137 how you use 139 requirements and best practices for 139 storage shelves commands for displaying information about 52 requirements for using multi-disk carrier 27 storage systems considerations for removing disks from 45 suggestions how to send feedback about documentation 199 SVMs assigning aggregates to 190 effect on aggregate selection 144 SVMs with Infinite Volume aggregate requirements 152 assigning aggregates 154 delegation 154 SyncMirror protection with RAID and 110 system capacity methods of calculating 13 system performance impact of media scrubbing on 24 T terminology family 147 third-party storage verifying back-end configuration 72 timeouts RAID option, considerations for changing 121 tips for creating and backing up aggregates, for sensitive data 21 topologies supported storage connection type 12 twitter how to receive automatic notification of documentation changes 199 U unmirrored aggregates explained 127 unprotected mode returning SEDs to 100 unreachable key management servers precautions taken in the event of during boot process 94 used space how to determine and control in aggregates, by volume 150 how to determine in aggregate 148 V V-Series functionality name change to FlexArray Virtualization 64 V-Series systems See LUNs (array) verifying back-end configuration 72 key management server links 91 vetoes of an aggregate relocation 188 overriding 188 VMDK drives how Data ONTAP reports disk types 10 volume command for displaying space information 53 volume show-footprint command understanding output 150 volumes creating aggregates for FlexVol, using physical drives 159 creating aggregates for Infinite, using physical drives 159 determining which ones reside on an aggregate 192 determining write-caching eligibility for Flash Pool aggregates 178 Index | 211
  • 212. how to determine space usage of, in aggregates 150 sharing of aggregates 152 See also Infinite Volumes Vservers See SVMs W WAFL external cache compared with Flash Pool aggregates 133 wear life how Data ONTAP manages SSD wear 26 wildcard characters how you use with disk ownership commands 62 wizards running the Storage Encryption setup 88 write caching determining Flash Pool aggregate eligibility 178 determining FlexVol volume eligibility 178 Z zoned checksum type changing for array LUNs 76 212 | Physical Storage Management Guide