How to Upgrade Physical Standby 2 Node RAC
My current environment is (all Oracle Enterprise Linux 5):
Two-node Oracle EE 11.1.0.7 RAC running on Clusterware 11.1.0.7 for my Primary database
Two-node Oracle EE 11.1.0.7 RAC running on Clusterware 11.1.0.7 for my Physical Standby database
I am attempting to upgrade Oracle EE to 11.2.0.2 and Clusterware to Grid Infrastructure 11.2.0.2
I only have one sticking point which I'm not sure of. My plan is to perform the following steps (Note: downtime is not an issue):
1. Stop log shipping and shutdown Standby RAC cluster.
2. Upgrade Standby nodes Clusterware to Grid Infrastructure 11.2.0.2
3. Startup Physical Standby. Database will still be 11.1.0.7 but with GI 11.2.0.2
4. Shutdown Primary RAC database
5. Upgrade Primary nodes Clusterware to Grid Infrastructure 11.2.0.2
6. Startup Primary RAC Database. Database will still be 11.1.0.7 but with GI 11.2.0.2
7. Resume log shipping
At this point, Grid Infrastructure will be 11.2.0.2 but Oracle EE will still be 11.1.0.7. I have had no problems with doing this to this point. At a later date, I will be upgrading Oracle EE to 11.2.0.2. My plan is:
1. Shutdown Primary database
2. Shutdown Standby database
3. Startup MOUNT Standby database under 11.2.0.2 ORACLE_HOME
4. Use DBCA to upgrade Primary to 11.2.0.2
a. Log apply will upgrade Standby as well
My problem is that I cannot use srvctl to start the Standby. I have to startup MOUNT the Standby with SQL*Plus. That part is all well and good and the upgrade goes fine. But after the upgrade, I cannot get the Standby back into a clustered database under Grid Infrastructure's control. If I try to start the Standby with srvctl, the OCR still thinks the database should be running under the old 11.1 home. How do I tell OCR to use the new 11.2 home for the Standby database?
I've poured through the documentation and Metalink but I haven't seen any good suggestions. There isn't much out there which tells me how to upgrade a RAC physical standby.
Any help is appreciated.
Cheers,
Brian