This leads to a peculiar scenario. A production database crashes due to a known bug. DBAs log severity 1 SR with Oracle. Oracle gives them a one off patch which is already available because this is a known bug. DBAs apply the patch, and the issue is fixed.
Business is aghast at this approach. Its like you wait for the instance to crash to apply a patch. They do not like this reactive practice at all. They want the DBAs to apply patches for all such known bugs. Oracle says that if you apply all these one off patches, you might end up with new bugs. So the path of least resistance is to apply those DB patches which are marked as Recommended in metalink.
This is a never ending story. Even if you upgrade to 10.2.0.4, there are bugs on top of 10.2.0.4 too. I am unsure if we can ever get out of this loop.
No comments:
Post a Comment