Unix 6 en

UNIX OS
Lecture VI
Simonas Kareiva
Vilnius University
Faculty of Mathematics
and Informatics
Preparation of the material was supported by the project „Increasing Internationality in Study
Programs of the Department of Computer Science II“, project number VP1–2.2–ŠMM-07-K-
02-070, funded by The European Social Fund Agency and the Government of Lithuania.

Lecture VI outline
File systems
Disks
NFS protocol
RAID arrays
2

Part I : file systems and disks
disk's structure and terminology
file system types
Linux file systems
indexing
3

a very typical hard disk
terms:
A) Track
B) Geometrical sector
C) Sector
D) Cluster
4

Clusters and sectors
Sectors make clusters
Size of a sector – 512 bytes, 2048 bytes
Sectors are grouped into clusters in order to save
addressing and number of disk requests.
Cluster is the smallest logical disk’s element which can
store information, so:
 5KB file in disk uses 8KB when cluster size is 4KB
5

FAT…
FAT… bits tell the cluster address table size:
FAT12 – maximum partition Size– 32MB 
 Max number of clusters – 2^12-12, clusters up to 8K
FAT16 – max partition size – ??? GB (exercise)
 Clusters up to 64K
FAT32 - (many nice features) – max partition still 2TB 
6

More file systems:
NTFS
UFS
EXT2
EXT3
ReiserFS
HFS
7

Disk devices (Linux):
/dev/
 /dev/hda
 /dev/hda1
 /dev/hda2
 /dev/hdb
 /dev/hdc
 /dev/sda
 /dev/sda1
 …
8

Disk devices (FreeBSD)
/dev/
 /dev/ad0
 /dev/ad0s1
 /dev/ad0s1a
 /dev/ad0s1b
 /dev/ad0s1c
 /dev/ad1
 /dev/ad1s1
 /dev/ad1s1a
 /dev/da0
 /dev/da0s1
 /dev/da0s1a
 …
9

Disk devices (Solaris)
/dev/dsk/
 /dev/dsk/c1
 /dev/dsk/c1t0
 /dev/dsk/c1t0d0
 /dev/dsk/c1t0d0p0
 /dev/dsk/c1t0d0s1
c – controller (control device)
t – SCSI target
d – LUN-id (mostly 0)
s – Solaris slice or p – primary partition
10

Working with disks (Linux)
# fdisk -l /dev/hdc
Disk /dev/hdc: 64 heads, 63 sectors, 787 cylinders
Units = cylinders of 4032 * 512 bytes
Device Boot Start End Blocks Id System
/dev/hdc1 * 1 610 1229728+ 83 Linux
/dev/hdc2 611 787 356832 5 Extended
/dev/hdc5 611 787 356800+ c Win95 FAT32 (LBA)
11

Creating a new disk
Create a partition with fdisk (RTM)
Linux:
 mkfs -t ext2 /dev/hdc1
 or
 mkfs.reiserfs /dev/hdc1
FreeBSD:
 newfs /dev/da0s1a
12

Mounting
mkdir /mnt/new
mount /dev/hdc1 /mnt/new
13

/etc/fstab
# FreeBSD fstab:
# Device Mountpoint FStype Options Dump Pass#
/dev/da0s1b none swap sw 0 0
/dev/da0s1a / ufs rw 1 1
/dev/da0s1e /tmp ufs rw 2 2
/dev/da0s1f /usr ufs rw 2 2
/dev/da0s1d /var ufs rw 2 2
/dev/da1s1a /squid ufs rw 2 2
/dev/acd0 /cdrom cd9660 ro,noauto 0 0
# Linux fstab:
/dev/rootvg/lv00 / ext3 defaults 1 1
/dev/rootvg/lv02 /vz ext3 defaults 1 2
/dev/md0 /boot ext3 defaults 1 2
tmpfs /dev/shm tmpfs defaults 0 0
devpts /dev/pts devpts gid=5,mode=620 0 0
sysfs /sys sysfs defaults 0 0
proc /proc proc defaults 0 0
/dev/rootvg/lv01 swap swap defaults 0 0
14

NFS protocol
Like Windows File Sharing, but not exactly...
Designed for intensive data exchange between servers
Easily configurable
From an OS point of view, its just another file system
(partition)
15

What you’ll need:
FreeBSD:
 nfsd – daemon, processing user requests
 mountd – daemon, responsible for FS mounting
 rpcbind – additional service used for chatting within the
network
/etc/exports:
 Format:
 /folder/which/shared –paremeters client1 client2…
16

/etc/exports:
/cdrom -ro host1 host2 host3
/home -alldands 10.0.0.2 10.0.0.3 10.0.0.4
/a -maproot=root host.example.com box.example.org
/usr/src /usr/ports host4
Then we execute:
# rpcbind
# nfsd -u -t -n 4
# mountd -r
17

Or use:
# showmount -e server
Exports list on server:
/usr 10.10.10.0
/a 10.10.10.0
# mkdand /mnt/a
# mount server:/a /mnt/a
# df -F nfs
Filesystem Type blocks use avail %use Mounted on
server:/a nfs 68510 55804 12706 81% /mnt/a
18

Practical uses
Most OS’s support installation through NFS, its enough to
have 1 DVD which is shared via NFS on LAN.
By using NFS you can create a couple of HTTP servers
with the same content and distribute traffic between
them by using the round-robin load balancing algorithm.
In large networks, it is useful to keep /home folders on
NFS, so that user’s files are saved centrally.
Exercise: think of more practical NFS uses!
19

RAID
terms:
 RAID – Redundant array of Inexpensive Disks
 JBOD – Just a bunch of … ?
 Hot spare – reserve player
 Stripe – data distribution in multiple devices
 Mirror – data backup in multiple devices
 Parity – protection from errors system (XOR):
 0 XOR 0 = 0
 0 XOR 1 = 1
 1 XOR 0 = 1
 1 XOR 1 = 0
 Dedicated parity – data in separate disk
20

RAID types
RAID0, RAID1, RAID2, RAID3, RAID4, RAID5, RAID6
RAID0+1, RAID1+0, RAID0+3, RAID51, …
21

Say,
We have four disks:
 Three 200GB disks
 One 300GB disk
22

RAID0
Size = n * min ( 200GB, 300GB ) = 800GB
Advantages:
 Fast reading
 Fast writing
 Lots of space
Disadvantages:
 Totally unreliable (if one of the disks
breaks down, the entire RAID stops)
23

RAID1
Size= min ( 200GB, 300GB ) = 200GB
Advantages:
 Fast reading
 Reliability and security
Disadvantages:
 Expensive writing (to both disks)
24

RAID5
Size= n-1 * min ( 200GB, 300GB ) = 200GB
Advantages:
 Fast reading
 Fast writing
 Reliability and security (parity)
Disadvantages:
 Requires at least 3 disks, often requires +1 additional
 Slow, when at least stripe size is being written
25

Parity
disk #1: -------- (Data)
disk #2: -------- (Data)
disk #3: -------- (Data)
disk #4: -------- (Data)
disk #5: -------- (Hot Spare)
disk #6: -------- (Parity)
26

Parity
disk #1: 00101010 (Data)
disk #2: 10001110 (Data)
disk #3: 11110111 (Data)
disk #4: 10110101 (Data)
disk #6: -------- (Parity)
00101010 XOR 10001110 XOR 11110111 XOR 10110101 = 11100110
D1 XOR D2 XOR D3 XOR D4 = ( (D1 XOR D2) XOR D3) XOR D4
27

Parity
disk #1: 00101010 (Data)
disk #2: 10001110 (Data)
disk #3: 11110111 (Data)
disk #4: 10110101 (Data)
disk #6: 11100110 (Parity)
28

Parity
disk #1: 00101010 (Data)
disk #2: 10001110 (Data)
disk #3: --Dead-- (Data)
disk #4: 10110101 (Data)
disk #5: 11110111 (Hot Spare)
disk #6: 11100110 (Parity)
00101010 XOR 10001110 XOR 11100110 XOR 10110101 = 11110111
29

Yo, dawg, I heard you like RAID…
… so we put more RAID into
your RAID so that you can have
more RAID!
33

Cheat-sheet of RAID configurations
35

Unix 6 en

More Related Content

What's hot (20)

Similar to Unix 6 en (20)

Unix 6 en