Re: SQL High version count because of too many varchar2 columns
From: Nigel Antell <nigel.antell_at_antelopeconsulting.co.uk>
Date: Mon, 17 Sep 2012 08:15:45 +0100
Message-Id: <555D814C-2930-4DDD-A4D8-1387D3CE1EB6_at_antelopeconsulting.co.uk>
Just to clear things up about the event 10503 - it _does_ work at session level. It was broken at session level in by another patch intending to fix the overall fact that it didn't work at all. I needed this event working at session level, so raised a bug (10274265) in MOS in Nov 2010. Oracle accepted that the event's intended functionality was indeed broken, as it is meant to work at session level. This took a long time to come through but was eventually patched under 11.2.0.3.0 on SPARC, and I see various other ports to AIX are now there.
Date: Mon, 17 Sep 2012 08:15:45 +0100
Message-Id: <555D814C-2930-4DDD-A4D8-1387D3CE1EB6_at_antelopeconsulting.co.uk>
Just to clear things up about the event 10503 - it _does_ work at session level. It was broken at session level in by another patch intending to fix the overall fact that it didn't work at all. I needed this event working at session level, so raised a bug (10274265) in MOS in Nov 2010. Oracle accepted that the event's intended functionality was indeed broken, as it is meant to work at session level. This took a long time to come through but was eventually patched under 11.2.0.3.0 on SPARC, and I see various other ports to AIX are now there.
We then set the event at session level under a login trigger for the application sessions suffering the issue. This removed the problem, and caused no bad side effects (in an environment with 60k SQL executions per second).
You may have to port the patch over to your specific version, but it is worth a try, particularly if the problem you are experiencing in QA is in a different part of the system.
Nigel Antell--
http://www.freelists.org/webpage/oracle-l
Received on Mon Sep 17 2012 - 02:15:45 CDT