OPatch bugs when applying Grid Infrastructure 12.1.0.2.5
By Martin | November 20th, 2015 | Category: 12cR1, Real Application Clusters | No Comments »For one of my clients, I experienced several issues with applying PSU 12.1.0.2.5 with opatch 12.1.0.1.9.
There were some code changes in opatchauto, which are not yet production-ready. I am looking forward to seeing a new opatch release (maybe 12.1.1.10 or 11) which has these issues fixed.
-
Bug 22091017 : OPATCHAUTO -ANALYZE COMMAND SHUTS DOWN THE RDBMS HOME IN 12C
“opatchauto apply -analyze” is normally used before patching to verify if there are any conflicts. In our case, the “opatchauto apply -analyze” did a shutdown of the database instances. Not good. - BUG 22202019 – OPATCHAUTO FAILS WITH “TOO MANY ARGUMENTS” FROM CHECKSPFFILE SCRIPT
When applying GI PSU patch to RDBMS Home and database instances are already stopped, patch apply fails. The reason is that opatchauto is trying to query database instances for information before shutting down instances. If instances are already stopped, these queries fail and opatchauto aborts. - OPatchauto is shutting down database instance with “srvctl stop home -stopoption TRANSACTIONAL”. I was of the opinion that this is doing a “shutdown transactional”, which in turn will wait until the last session is performing a commit. This can take a long time. Support Engineer explained that the wait will be limited to 10min, after which the system is begin stopped with “shutdown abort”. This is not documented in “http://docs.oracle.com/database/121/RACAD/srvctladmin.htm#RACAD5040” and in my opinion a bug. I expect the shutdown to be “immediate” whithout delaying the patching process.