In Oracle 11g, we can now do rolling upgrades of the ASM instanaces in a RAC cluster configuration. In this example shown below, one of the ASM instances is being upgraded to 11.1.0.7 while the other instance is still running on version 11.1.0.6. While the ASM software is being upgraded, client connections to the RAC database will continue as the service will get relocated to the other node in the cluster which is not being currently upgraded. So in other words this is an online patch application.
The database name here is racdba and the two instances are racdba1 and racdba2 running on nodes middba1 and middba2. The platform used in this example is AIX 5.3 64 bit on a P series LPAR.
On one of the nodes in the cluster, connect to the ASM instance (+ASM1) and issue the command to start the rolling migration.
SQL> alter system start rolling migration to '11.1.0.7'; System altered.
Stop the database instance racdba1, shutdown the ASM instance +ASM1 and then stop the listener which is running from the ASM home which will now be upgraded to 11.1.0.7.
middba1 $ srvctl stop instance -d racdba -i racdba1 middba1 $ sqlplus sys as sysasm SQL*Plus: Release 11.1.0.6.0 - Production on Fri Dec 11 09:57:13 2009 Copyright (c) 1982, 2007, Oracle. All rights reserved. Enter password: Connected to: Oracle Database 11g Enterprise Edition Release 11.1.0.6.0 - 64bit Production With the Partitioning, Real Application Clusters, OLAP, Data Mining and Real Application Testing options SQL> shutdown immediate; ASM diskgroups dismounted ASM instance shutdown SQL> quit middba1 $ lsnrctl stop LISTENER_MIDDBA1 LSNRCTL for IBM/AIX RISC System/6000: Version 11.1.0.6.0 - Production on 11-DEC-2009 10:00:31 Copyright (c) 1991, 2007, Oracle. All rights reserved. Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=middba1_vip)(PORT=1521)(IP=FIRST))) The command completed successfully
Initially the service ‘rac1_service’ is connected to instance racdba1 hosted on node middba1. When we shutdown the instance racdba1, the service will be relocated to the other surviving instance running on node middba2.
middba2 $ sqlplus sh/sh@rac1
SQL*Plus: Release 11.1.0.6.0 - Production on Fri Dec 11 09:11:44 2009
Copyright (c) 1982, 2007, Oracle. All rights reserved.
Connected to:
Oracle Database 11g Enterprise Edition Release 11.1.0.6.0 - 64bit Production
With the Partitioning, Real Application Clusters, OLAP, Data Mining
and Real Application Testing options
SQL> select count(*) from sales;
COUNT(*)
----------
918843
SQL> select host_name from v$instance;
HOST_NAME
----------------------------------------------------------------
middba1
********ASM and Database Instance is shutdown here ****************
SQL> /
HOST_NAME
----------------------------------------------------------------
middba2
SQL> select count(*) from sales;
COUNT(*)
----------
918843
While installing the 11.1.0.7 patch, it should be noted that we are only selecting one node in the cluster where the patch will be applied – the other node in the cluster is where the 11.1.0.6 ASM and Database instance is running and where all the client connections are being hosted.
After the 11.1.0.7 patch application on the ASM Oracle Home is completed, we will now start the listener, the upgraded ASM instance and the database instance.
middba1 $ lsnrctl start LISTENER_MIDDBA1 middba1 $ sqlplus sys as sysasm SQL*Plus: Release 11.1.0.7.0 - Production on Fri Dec 11 10:48:46 2009 Copyright (c) 1982, 2008, Oracle. All rights reserved. Enter password: Connected to an idle instance. SQL> startup ASM instance started Total System Global Area 283930624 bytes Fixed Size 2153056 bytes Variable Size 256611744 bytes ASM Cache 25165824 bytes ASM diskgroups mounted middba1 $ srvctl start instance -d racdba -i racdba1
We will now repeat all the same steps on the other node middba2. After the 11.1.0.7 patch has been applied to all the ASM Oracle Home’s we will see a message indicating that we now can turn off the ASM rolling upgrade.
Note: while the ASM in in ‘rolling upgrade’ mode, a limited number of ASM operations can be performed and certain ASM operations like rebalancing cannot be carried out.
SQL> alter system stop rolling migration; System altered. We can finally relocate the service back to its original node - middba1middba1 $ srvctl relocate service -d racdba -s rac1_service -i racdba2 -t racdba1 -f middba1 $ srvctl status service -d racdba -s rac1_service Service rac1_service is running on instance(s) racdba1
