RE: oracle workaroud vs patch

From: Hsieh, Joan <Joan.Hsieh_at_tufts.edu>
Date: Tue, 18 Aug 2015 12:26:25 +0000
Message-ID: <416FB269D9362D4AA40BE7124CA3EF9F3520F703_at_SSVMEXDAG01MB02.tufts.ad.tufts.edu>

Thanks, Mark. Your advice and insight is very much appreciated. Joan

From: Mark W. Farnham [mailto:mwf_at_rsiz.com] Sent: Monday, August 17, 2015 7:44 PM
To: Hsieh, Joan; oracle-l_at_freelists.org Subject: RE: oracle workaroud vs patch

The workaround means you will not get the advantage of the optimizer_adaptive_features which may from time to time reduce the actual run time cost of executing some queries the 2nd through nth time the instance is up. If you don't get enough advantage from these features that you know about them, then this is likely less destabilizing and less risky than applying a patch.

If you apply the patch, then you retain any advantage you are getting from optimizer_adaptive_features at the cost of my general claim that any patch has inherent risks greater than toggling a parameter to disable a performance feature.

If someone has actual experience with this particular patch they may be able to tell you whether or not there was any regression in Oracle behavior (or that the patch had significant side effects; let's hope for the former).

If you have a sandbox and a regression suite that tests all your relevant transactions and queries relevant to your production use for comparison, you might be able to quickly verify that the patch does not break anything important to you.

Good luck,

mwf

From: oracle-l-bounce_at_freelists.org<mailto:oracle-l-bounce_at_freelists.org> [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Hsieh, Joan Sent: Monday, August 17, 2015 11:36 AM
To: oracle-l_at_freelists.org<mailto:oracle-l_at_freelists.org> Subject: oracle workaroud vs patch

Hi,

 Our 12.1.0.2 database seems to hit the bug 18430870<https://support.oracle.com/rs?type=bug&id=18430870>, ADAPTIVE PLAN AND LEFT JOIN GIVE WRONG RESULT, there is a workaround to set the optimizer_adaptive_features=false, also, there is a patch that Oracle provides to fix the bug. I'd like to know what's pro's and con's for the 2 approaches. Should the work around is sufficient to deal with this bug?

 Thanks,

 Joan

--
http://www.freelists.org/webpage/oracle-l

Received on Tue Aug 18 2015 - 14:26:25 CEST