Cdb part i

Copyright © 2012, Oracle and/or its affiliates.All rights reserved.
<Insert Picture Here>
Oracle Database 12c –
Pluggable Databases – Part I

Using the Player
Course Outline
Player Controls
Print Version
Change View

Road Map

Road Map
PART – II
PART – I
PART – III

PROPERTIES
Allow user to leave interaction: After viewing all the steps
Show ‘Next Slide’ Button: Show upon completion
Completion Button Label: Next Slide

Why Take This Course?
• What’s in it for me?
• What are challenges I face on the job?
• How can this help me do my job better?

Basics of Container and Pluggable Databases

Road Map
PART – I

Challenges
Many Oracle customers have large numbers of “departmental”
applications built on Oracle RDBMS.
• Do NOT use a significant percentage of the hardware on
which they are deployed
• Have instance and storage overhead preventing large
numbers of “departmental” databases from being placed
on the same physical and storage server
• Are NOT sufficient complexity to require 100% of the
attention of a full time administrator
• Do require significant time to patch or upgrade all
applications

Benefits of 12c Pluggable Databases
• Operates multiple monolithic databases in a centrally
managed platform to lower costs:
– Less instance overhead
– Less storage cost
• Reduces DBA resources costs and maintains security
– No application changes
– Fast and easy provisioning
– Time saving for patching and upgrade
– Maintain separation of duties between:
– Different application administrators
– Application administrators and DBA
– Users within application
• Provides isolation

Benefits of 12c Pluggable Databases
• Ensures full backwards-compatibility with non-CDB
• Fully operates with RAC
• Is integrated with Enterprise and Resource Manager
• Allows central management and administration of
multiple databases
– Backups / Disaster recovery
– Patching and Upgrades

User Data is Added
In a non-CDB, user data is added:
• The metadata is mixed with the Oracle supplied data in the
data dictionary.
…
OBJ$ TAB$ SOURCE$
…
EMP DEPT
Oracle System data mixed with
User metadata
User Data

Separating System and User Data
…
OBJ$ TAB$ SOURCE$
…
EMP DEPT
User Metadata only User Data
…
OBJ$ TAB$ SOURCE$
Oracle Metadata only
Container Database
Container
for
System
Container for Application

Naming the Containers within Container Database
• The Oracle supplied
container is called the root
container
• User container is a
pluggable database (PDB)
Oracle Metadata only
root
Container Database
PDB
User Metadata only User Data

Provisioning Pluggable Databases
Seed
PDB
New
empty
PDB
Imagine this
when the datafiles
are on a filesystem
that supports
copy-on-change
root
Four methods:
• Create new PDB from PDB$SEED pluggable database
• Plug in a non-CDB
• Clone a PDB from another PDB into the same or another
CDB
• Plug an unplugged PDB into another CDB
Container Database

Container Database Architecture
Instance
System Global Area
Process Structures
Server
Instance
Container
database
root container
Datafiles Redo Log files
Control files
UNDO
TEMP
SYSTEM
SYSAUX
seed PDB
Datafiles
SYSTEM
SYSAUX EXAMPLE
SALES PDB
SYSTEM
SYSAUX TBS2
HR PDB
Data
files
TEMPO
Datafiles
TEMP2
PDBid2 PDBid3 PDBid4 PDBid2 PDBid4
Single DB shares:
• Background
processes
• Shared/process
memory
• Oracle metadata
• Redo log files
• Control files
• Some datafiles

Terminology
• Common vs Local:
– Users
– Roles
– Privileges
– Objects
• CDB vs PDB level:
– Global Resource Manager plan vs local RM plan
– Unified audit at CDB or PDB level
– Xstream at CDB or PDB level

Common and Local Users
Container Database
root
PDB_HR
PDB_SALES
Table
HR.REGIONS
USERNAME COMMON
SYS YES
SYSTEM YES
CDBA YES
HR YES
View CDB_USERS
Adding a common user, involves adding a
description of that user in the root and in every
PDB.
Certain operations, such as ALTER SESSION
SET CONTAINER can only be executed by
common users. Then depending on the way
the privileges are granted, common users can
or cannot perform operations.
A local user is a traditional user, known only
in its own PDB.
View DBA_USERS
USERNAME COMMON
SYS YES
SYSTEM YES
CDBA YES
HR YES
SALES NO
USERNAME COMMON
SYS YES
SYSTEM YES
CDBA YES
HR YES
HR_MGR NO
View DBA_USERS
Table
SALES.COUNTRIES
Tablespace
EXAMPLE
Tablespace
EXAMPLE
Tablespace
EXAMPLE

CDB_xxx All of the objects in the container database across all PDBs
USER_xxx Objects owned by the current user
ALL_xxx Objects accessible by the current user
DBA_xxx All of the objects in a container or pluggable database
Data Dictionary Views
• CDB_pdbs : All PDBS within CDB
• CDB_tablespaces : All tablespaces within CDB
• CDB_users : All users within CDB (common and local)
DBA dictionary views providing information within PDB:
SQL> SELECT view_name FROM dba_views WHERE view_name like 'CDB%';
SQL> SELECT table_name FROM dict WHERE table_name like 'DBA%';

Backup
• Whole CDB backup: control files, root and PDBs
datafiles
• Whole PDB backup: PDBs datafiles
• Hot backups: not available in Beta1
– PDB level
Backup/Recovery/Flashback Database

Recovery
• Instance recovery: CDB level only
• Complete media recovery from file loss or damage
– CDB level
– PDB level: not available in Beta1
– Tablespace level: same as for non-CDB
• Incomplete media recovery after file loss or corruption
– CDB level: same as for non-CDB
– PDB level not available in Beta1
– TSPITR for root tablespaces ONLY except SYSTEM,
UNDO, SYSAUX
• Flashback database
– CDB level only
– PDB level not available in Beta1
• Block recovery: no change
Backup/Recovery/Flashback Database

Migrating pre-12c Databases to 12c CDB
Container Database CDB1
root
Data files Redo Log
files
Control
files
Create
PDB1
from
ORCL
ORCL
Datafiles Control
files
Redo Log
files
expdp
Dump file
PDB1
Data files
impdp
There are two methods:
Method 1
Upgrade a pre-12c database to 12c
Plug-in the non-CDB into a CDB
using different methods
Or
Method 2
1. Pre-create a PDB in CDB
2. Use 11g expdp / 12c impdp
or
2. Use replication between the non-CDB
and the PDB
ORCL
Datafiles Control
files
Redo Log
files
1
2
11g
12c
2
1
Replication
DBMS_PDB
XML file

• One character set for all PDBs (Unicode recommended)
• PDB initialization parameters but a single SPFILE
• No PDB qualified database object names
– SELECT * FROM HR:apps.tab1
– Use DB Links: SELECT * FROM apps.tab1@HR
• Data Guard at CDB level
• Database Vault per PDB only
• Unified audit both at CDB and PDB level
Impacts

PROPERTIES
Allow user to leave interaction: Anytime
Show ‘Next Slide’ Button: Show always
Completion Button Label: Next Slide

Lesson Summary
In this first lesson, we discussed:
• The benefits for a container database
• The architecture of a container database
• The root container and the pluggable database container
• The methods for provisioning new pluggable databases
• The backup and recovery capabilities and upgrade
methods
• The impacts
PART – I

PROPERTIES
On passing, 'Finish' button: Goes to Next Slide
On failing, 'Finish' button: Goes to Next Slide
Allow user to leave quiz: At any time
User may view slides after quiz: At any time

Configuring and Creating a Container Database

Road Map
PART – I

Create a container database:
• To consolidate many pre-12.1 non-CDBs into a single,
larger database
• To prepare a container
– For plugging any future new application
– For testing applications
– For diagnosing application performance
– For unplugging/plugging operations
• To simplify and reduce time for patching and upgrade
Goals

Instance
SGA
Process Structures
Instance
Container
Database
CDB1
root container
seed pluggable database
Steps to Create a Container Database
2
1
Datafiles
Control
files
UNDO
SYSTEM
SYSAUX
Redo Log
files
SYSTEM
TEMP1
TEMP
SYSAUX
initCDB1.ora
3
4
Execute
catalog.sql
catproc.sql

A CDB has new characteristics compared to non-CDBs:
• Two containers:
– The root (CDB$ROOT)
– The seed PDB (PDB$SEED)
• Several services: one per container
– Name of root service = name of the CDB: cdb1
• Common users in root and seed: SYS,SYSTEM …
• Common privileges granted to common users
• Common roles
• Tablespaces and datafiles associated to each container:
– root: SYSTEM contains system-supplied metadata and no user
data, SYSAUX
– seed: SYSTEM, SYSAUX
After CDB Creation – What’s New in CDB

CDB_xxx All of the objects in the container database (new column CON_ID)
USER_xxx Objects owned by the current user in a PDB
ALL_xxx Objects accessible by the current user in a PDB
DBA_xxx All of the objects in the root or a pluggable database
Data Dictionary Views – CDB_xxx and V$xxx
CDB dictionary views provide information across PDBs:
• CDB_pdbs : all PDBS within the CDB
• CDB_tablespaces : all tablespaces within the CDB
• CDB_data_files : all datafiles within the CDB
• CDB_users : all users within the CDB (common and local)
SQL> SELECT view_name FROM dba_views
2 WHERE view_name like 'CDB%';

In this second lesson, we discussed:
• Create a container database
• Explore CDB instance, files and PDB structures
Lesson Summary
PART – I

Pluggable Databases Provisioning

Click to edit Master title style
Goals
Why provisioning pluggable databases in a CDB ?
• To consolidate many non-CDBs into a single, larger
database, the CDB
– To lower costs
– To better exploit hardware and DBA resources
• To avoid the creation of new databases and instances
when:
– Implementing new application
– Testing new applications
– Diagnosing new application performance
• To simplify and reduce time for patching and upgrade

Click to edit Master title style
Provisioning New Pluggable Databases
Four methods:
• Create a new PDB from the seed PDB
• Plug a non-CDB in a CDB
• Clone a PDB from another PDB:
– Into the same CDB
– Into another CDB
• Plug an unplugged PDB into another CDB

Method 1 – Create New PDB from PDB$SEED
root
Data files/
Tempfiles
Redo Log
files
Control
files
UNDO
TEMP
PDB$SEED
Data files
SYSAUX
USERS
SYSTEM
SYSAUX
PDB1
Data files
Create
PDB1
from
PDB$SEED
• Copies the datafiles from
PDB$SEED datafiles
• Creates tablespaces SYSTEM,
SYSAUX
• Creates a full catalog including
metadata pointing to Oracle
supplied objects
• Creates common users:
– Superuser SYS
– SYSTEM
• Optionally creates a local user
granted local PDB_DBA role
• Creates a new default service
SYSAUX
SYSTEM
SYSTEM

Steps – With FILE_NAME_CONVERT
Steps to create a new PDB from the seed using
FILE_NAME_CONVERT clause:
1. Connect to the root as a common user with CREATE
PLUGGABLE DATABASE system privilege:
2. Use views to verify:
SQL> CREATE PLUGGABLE DATABASE pdb1
2 ADMIN USER admin1 IDENTIFIED BY p1 ROLES=(CONNECT)
3 FILE_NAME_CONVERT = (’PDB$SEEDdir’,’PDB1dir’);
SQL> CONNECT / AS SYSDBA
SQL> SELECT * FROM cdb_pdbs;
SQL> SELECT * FROM cdb_tablespaces;
SQL> SELECT * FROM cdb_data_files;
SQL> ALTER PLUGGABLE DATABASE pdb1 OPEN;
SQL> CONNECT sys/password@pdb1 AS SYSDBA
SQL> CONNECT admin1/p1@pdb1

Method 2 – Clone PDBs within same CDB
root
Data files/
Tempfiles
Redo Log
files
Control
files
PDB$SEED
Data files / Tempfiles
PDB1
Data files
Create
PDB3
from
PDB1
1. In init.ora, set DB_CREATE_FILE_DEST=
'PDB3dir' or
PDB_FILE_NAME_CONVERT='PDB1dir,
PDB3dir'
2. Connect to the root as a common user
with CREATE PLUGGABLE DATABASE
privilege
3. Quiesce PDB1: Close PDB1 first, then
4. Clone PDB3 from PDB1:
5. Open PDB3:
PDB3
Data files
PDB3 owns:
• SYSTEM, SYSAUX tablespaces
• Full catalog
• SYS, SYSTEM common users:
• Same local administrator name
• New service name
SQL> ALTER PLUGGABLE DATABASE
2 pdb1 OPEN READ ONLY;
SQL> CREATE PLUGGABLE DATABASE
2 pdb3 FROM pdb1;
2 pdb3 OPEN;

Migrating pre-12.1 Databases to 12.1 CDB
There are two methods:
1. Upgrade an existing pre-12.1 database to 12c
2. Plug-in non-CDB into a CDB
Or
1. Pre-create a PDB in CDB
2. Use 11g expdp / 12c impdp
or
2. Use replication between non-CDB and PDB

Migrating pre-12.1 Databases to 12.1 CDB
First method:
1. Upgrade an existing 11g database to
12c
2. Plug-in the non-CDB into a CDB
using:
• Export/import
– TTS
– TDB
– Without transportable
• Replication
• DBMS_PDB.DESCRIBE to
generate XML file
root
Data files/
Tempfiles
Redo Log
files
Control
files
Create PDB1
from
ORCL
ORCL
Datafiles Control
files
Redo Log
files
expdp
Dump file
PDB1
Data files
impdp
ORCL
Datafiles Control
files
Redo Log
files
1
2
11g
12c
Replication XML file

Dropping a PDB
root
Data files/
Tempfiles
Redo Log
files
Control
files
File2
TEMP
PDB$SEED
Data files
File5
File6 File7
File4
File1
File3
PDB1
Data files
File8 Drop PDB1
• If KEEP DATAFILES (default):
– Retain datafiles
– Can be plugged in another or
same CDB
• If INCLUDING DATAFILES:
– Removes PDB1 datafiles
• Updates controlfiles
• Cannot drop seed PDB
File9 File10
PDB4
Data files
2 pdb1 CLOSE;
SQL> DROP PLUGGABLE DATABASE
2 pdb1 [INCLUDING DATAFILES];
…
File7
File8
File9
File10
…

Lesson Summary
In this third lesson, we discussed:
• The different methods to provision new pluggable
databases
• The migration methods
• The impact of dropping a pluggable database
PART – I

Course Summary
In this first part of the course, we discussed:
PART – II
PART – I
PART – III
PART – II

Cdb part i

More Related Content

What's hot (20)

Similar to Cdb part i (20)

Recently uploaded (20)

Cdb part i