MSWIn, 9i -> 11G Upgrade, 100% CPU Usage, No Workload
Date: Tue, 16 Apr 2013 09:08:03 -0700 (PDT)
Message-ID: <dea5aee4-7f30-4b36-a034-45f3b42071b4_at_googlegroups.com>
Hi All,
I have been running fully patched 11G instances in my development and staging environments, MSWin2003, for over 2 years, never an issue.
Finally, we do the production upgrade with a MSWin2008 on HyperV and within 30 minutes CPU is at 100% with only 2 connections made and no workload, nothing current in v$sql*. Rollback......
Repurpose my hot standby (MSWin2003, bare metal), temporarily alias it as the production database and the same problem occurs. After a physical reboot of the server, issue re-occurs within 10 minutes.
Oracle Support says turn off all jogs, particularly the DEFAULT_MANAGEMENT_PLAN (?). Repeat the test problem occurs. I believed I had identified the cause, possible bad statistics in a schema.
Move the machine to a test environment with the identified client. Test the client with a connection and simple report to my staging database (STG), which immediately consumes 100% CPU. I test the client against my "prod now test" databases, with an exact copy of what had been done to cause the problem in production. Manually, create the db, datafiles and users, import from 9I. Save and calculate stats at various different times. However, I cannot consistently reproduce the problem. I am successful on 2 out 10 tries.
All occurrences or the problem on new databases will repeat almost immediately after a bounce, STG will stay up for a day or two or more before it occurs again.
STG has had lots of drop and create users and imports but the database itself is over 1 year old.
We do not run OEM or emctl, I'm a command line kind of guy.
I'm baffled, any ideas? I have opened an SR but .....
TIA
Dave
Received on Tue Apr 16 2013 - 18:08:03 CEST