Você está na página 1de 8

Oracle Database Upgrade to 11gR2

From ETWiki
(Redirected from SOP - Upgrading 10G,11g database's to 11.2.0.2 version)

Contents
[hide]

· 1 Check List

· 2 Phase I: Pre Upgrade Tasks

· 3 Phase II: Upgrading to the New Oracle Database 11gR2

· 4 Phase III: Post-Install Oracle Database 11gR2

· 5 Issue or knows error while upgrade

[edit] Check List


Scope : Wiki is describing pre/post checklist and upgrade process details for upgrading a
Oracle 10R2 or 11.2.0.1 to Oracle 11gR2(11.2.0.2)

· Please check OS version as upgrade is only certified for RH 5.


cat /etc/red*

· Please ensure that latest package is pushed by Onsite Oracle DBA team.
DBA ADMIN Team Part: update the EM files (oratab, tnsnames.ora, listener.ora). Push the
package from adm0m3:
--For Linux: On Target Server install latest Packages : --
etrpmadd -f -p oracle -v 11.2.0.2.x86_64
etrpmadd -f -p oradba -v 1.0
etrpmadd -f -p oradiag -v 1.0
--FROM adm0m3
etupdate –fn <server_name>
etupdate –fn <server_name> -p /etrade/lib
etupdate –fn <server_name> -p /etrade/bin

· Verify all ET standards


* Index Placement
* DB User profile assignment
* Physical Naming convention.
* File Naming conventions.

· Etrade DBA Task:


* Install latest 11202 package
* Execute etserver install on new envirment so ORAPWD and init links will be
created on new ORACLE_HOME
* etcmd -s chmod 777 /etrade/etc/packages.linux.intel on target Server
Phase 1: Pre Upgrade Tasks.
Phase 2: Upgrading to Oracle Database 11g Release 2
Phase 3: Post Upgrade Tasks.

[edit] Phase I: Pre Upgrade Tasks


· Step 1: Pre-create DIR.
Create upgrade directory at $ET_INSTANCE_ROOT/bkup/ location
etcmd mkdir upgrade pre_upgrade_bkup

· Step 2 :
Disable all batch and cron jobs.
Backup
Archive removal and rotation
Application jobs
Stats collection job
Netcheck Monitoring. (In Oracle CRON)
standby job at standby server and make sure all arch has been applied at DR/STANDBY
server side.

· Step 4 : Gather full stats for the database


etcmd gather_stats.ksh db
etcmd gather_stats.ksh full

· Step 6: Capture Database link information.


Create a list of database link before starting an upgrade as can be useful many time like in case
of downgrade or in post task after upgrade.
SQL> spool create_dblink.info
SQL> SELECT
'create '||DECODE(U.NAME,'PUBLIC','public ')||'database link '||CHR(10)
||DECODE(U.NAME,'PUBLIC',Null, U.NAME||'.')|| L.NAME||chr(10)
||'connect to ' || L.USERID || ' identified by
||L.PASSWORD|| using || L.host || '
||chr(10)||';' TEXT
FROM sys.link$ L, sys.user$ U WHERE L.OWNER# = U.USER# ;
SQL> spool off

· Step 7: Create Pre-Upgrade BackupSET.


Backup the existing database or Please ensure for we do have a latest backup for the database.
etcmd dba_bkup_gzip.ksh

Note : --> Provide rac as first argument if it is a Rac enviroment.


IMP: Copy backup to either standby server or move it to pre_upgrade_bkup
folder

· Step 8: Verify Disk free space.


Check for adequate free space on archive log destination file systems.
[plt:etadm:dbs:PORA11G:pltdb2w98m7] df -h | grep arch
/dev/vx/dsk/pltdb2w98m7dg/arch
33G 20G 13G 62% /etrade/plt/etadm/dbs/PORA11G/arch

Create HD to Increase space if needed. Ideally it should be 10~15G free.

· Step 9: Ensure no files need media recovery:


SQL> select * from v$recover_file;
no rows selected

Make sure currently no backup is running for the database, if yes then end
backup using command.

SQL > Alter database end backup;


Database altered.

· Step 10 : Verify DB version and installed options.


Ensure the properties for SYSTEM ,SYS user and database component status should be valid
and default tablespace should be SYSTEM and SYSAUX respectively.
SQL> select username, default_tablespace from dba_users where username in
('SYS','SYSTEM');
==============================================================================
===============
SYS SYSTEM
SYSTEM SYSTEM

2 rows selected.

col COMP_NAME for a39


col VERSION for 9999999
col status for a19
set lines 1000
select COMP_NAME, VERSION, STATUS FROM SYS.DBA_REGISTRY;

=============================================================================
Oracle Database Catalog Views 11.2.0.1.0 VALID
Oracle Database Packages and Types 11.2.0.1.0 VALID

Note : If any component is invalid then you should work bringing them in valid
status to procees ahead.

· Step 11: Make sure Recycle bin is empty.


Empty the Recycle Bin using purge statement.
SQL> PURGE DBA_RECYCLEBIN;
DBA Recyclebin purged.

· Step 12: Recompile invalid objects:


SQL> spool invalid_objects_pre_upgrade_11g.log
SQL> @${ET_ORADBA_ROOT}/sql/list_invalid_objects.sql
SQL> spool off

Note : Compile all objects before proceeding ahead.

· Step 13 Spool file before running upgrade tool


SQL> spool pre_upgrade_database_info.log
SQL> @/etrade/pkgs/linux/intel/oracle/11.2.0.2.x86_64/rdbms/admin/utlu112i.sql
SQL> spool off

Then, check the spool file and examine the output genrated by utlu102i.sql.

· Step 14: Shutdown Database.


Stop the database cleanly and also stop all applicable tools running for like Rember: You need to
shutdown database by using older ORACLE_HOME. To shutdown the database connect as
etcmd sqlplus " / as sysdba"
SQL> shutdown immediate;

To stop agent running on host.


$ etcmd -a oracle emctl stop agent

To stop listner on the database :--


$ etlsnrctl stop
Imp : -- For a cluster database, set the parameter CLUSTER_DATABASE=FALSE during the
upgrade. Upgrading a cluster database, then modify the initdb_name.ora file in the same way that
you modified the parameter file.

[edit] Phase II: Upgrading to the New Oracle Database


11gR2
· Step 1 : Update ET Framework Files:
Make required changes in file to have latest environment value's to take affect and re login from
a new session NOTE: Keep your older PUTTY window with older ORACLE_HOME alive.
vi /etrade/etc/packages.linux.intel

Grep the specific database entry and change the value to latest home.

Like for PORA11G it was

#plt:etadm:dbs:PORA11G:oracle:11.2.0.1.x86_64 --- OLD value


plt:etadm:dbs:PORA11G:oracle:11.2.0.2.x86_64 --- New value after changes

· Step 2: Source new Oracle home variable by re login


Make Sure that ORACLE_HOME variable is pointing to new ORACLE_HOME
--creating new password file under new oracle 11g Home : --
$etcmd orapwd file=${ORACLE_HOME}/dbs/orapw${ORACLE_SID} password={password}
entries =80 force=y

· Step3: Creating links and update PFILE

· Create pfile and orapwd file links by running etserver install.

· Rename Old parameter file present in ${ET_INSTANCE_ROOT}/admin/pfile/ to


*.10g.

· Copy sample File from ${ET_ORADBA_ROOT}/dbcreate/11g/initSAMPLE11G.ora


to ${ET_INSTANCE_ROOT}/admin/pfile/init{ORACLE_SID}.ora

· Take care of local_listener parameter

· Make appropriate changes in pfile


Make compatible parameter to 11.2.0.1. This will allow us to easy downgrade.
compatible will change only after testing and confirmation.

###########################################
# Miscellaneous
###########################################
compatible = 11.2.0.1

· It is Assumed that database was shutdown cleanly using OLD Oracle_HOME at end
of Phase 1.
etcmd sqlplus " / as sysdba"
SQL> startup nomount;
SQL> shutdown

· Step 4 : Startup database in upgrade mode and run the upgrade script :
This time you will be login to database using 11202 ORACLE_HOME.
SQL*Plus: Release 11.2.0.2.0 Production on Tue Feb 1 14:00:11 2011
Copyright (c) 1982, 2010, Oracle. All rights reserved.
Enter password:
Connected to an idle instance.
SQL*Plus: Release 11.2.0.2.0 Production on Tue Feb 1 14:00:11 2011
Copyright (c) 1982, 2010, Oracle. All rights reserved.

Enter password:
Connected to an idle instance.
SQL> spool $ET_INSTANCE_ROOT/bkup/upgrade/upgrade_to_11_2_0_2.log append
SQL> startup upgrade
SQL> @?/rdbms/admin/catupgrd.sql

Imp : -This script will take care of all required script execution for upgrade existing installed
packages.
Oracle Database Catalog Views
Oracle Database Packages and Types
Oracle Real Application Clusters -- For RAC database's.

SQL> SPOOL OFF


SQL> startup

Check the spool file and verify that the packages and procedures compiled successfully. Correct
any problems you find in this file and rerun the appropriate upgrade script if necessary. You can
rerun any of the scripts described in this note as many times as necessary.
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

DBA ADMIN ORACLE TEAM TASK : --- Ask to update all required file at admin
server adm0m3
Packages.linux.intel, oratab, tnsnames.ora, listener.ora
and Push to target server

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

[edit] Phase III: Post-Install Oracle Database 11gR2


· Step 1: Verify Upgrade Status:
Executing Oracle Script : utlu112s.sql This is the Post-upgrade Status Tool displays the status of
the database components in the upgraded database. The Upgrade Status Tool displays output
similar to the following:
SQL> @?/rdbms/admin/utlu112s.sql

.
Oracle Database 11.2 Post-Upgrade Status Tool 01-28-2011 10:45:32
.
Component Status Version HH:MM:SS
.
Oracle Server
. VALID 11.2.0.2.0 00:10:43
Gathering Statistics
. 00:00:40
Total Upgrade Time: 00:12:03

PL/SQL procedure successfully completed.

Note - in RAC environments, this script may suggest that the status of the RAC
component is
INVALID when in actual fact it is VALID (as shown in the output from
DBA_REGISTRY)

Verify the DATABASE components


SQL> select comp_id, version, status from DBA_registry;
COMP_ID VERSION STATUS
------------------------------ ------------------------------ ------------
CATALOG 11.2.0.2.0 VALID
CATPROC 11.2.0.2.0 VALID
Verify DST Information: Oracle 11gR2 use the DST zone file of 11 (or higher). After upgrade
you should see
SQL> SELECT * FROM v$timezone_file;

FILENAME VERSION
-------------------- ----------
timezlrg_11.dat 11

· Step 3: Verify Invalid Objects:


Check Invalid objects and compare with details of objects at PRE-UPGRADE file
invalid_objects_pre_upgrade_11g.log status.
SQL> shutdown immediate (DO NOT USE SHUTDOWN ABORT!!!!!!!!!)

SQL> startup;
SQL> spool invalid_objects_post_upgrade_11g.log
SQL> @${ET_ORADBA_ROOT}/sql/list_invalid_objects.sql
SQL> spool off

· Step 4: Start the database,agent,listner cleanly:


Reduce Memory parameters of POOL sizes which increased for upgrade." If any "
SQL> shutdown immediate
SQL> startup

Start the listener and OEM if applicable.


$ etlsnrctl start
$

· Step 5: Verify connectivity:


Verify connectivity to new database using all available tnsnames.
$ sqlplus etdbmon/<pwd>@TNS_ENTRY

· Step 7: Gather full Stats


$ etcmd gather_stats.ksh full

· Step 9: Enable cron and batch jobs:


1) Backup 2) Archive removal and rotation 3) Application jobs 4) Stats collection job 5) Net
check Monitoring.

· Step 10 : Verify ET Standards.


Verify soft links
Correct UNDO/TEMP/Redo

· Step 11: Enable flashback:

Você também pode gostar