My company recently had an assessment performed on our Oracle DB, and quite a few recommendations were passed back to us. As someone who knows very, very little about Oracle, I'm trying to assess whether or not these statements are as dire as they sound. (We do have an Oracle DBA, and it's not me ). Obviously, with a single DBA, this isn't a huge system, but I'm not sure what metrics to give here for reference.
It seems that there has not been a regular performance analysis of database automated workload repository and diagnostics monitor. No daily checklist of monitoring and management activities. Mostly reactive care, rather than proactive.
Tests done during this assessment showed that there was a pattern of virtual memory being exhausted. This was not discovered prior to the assessment.
The File System IO was not set to direct IO in the initialization parameter file.
Installed release is 10gR2, which means we are 7 releases behind, and have missed over 25 Critical Patch Updates.
So - are these signs of a distracted/overburdened DBA? Does this sound like a normal system? Any thoughts you guys may have would be greatly appreciated.
10gr2 is not 7 releases behind, its 2 - 11g r2 has only just been released so you are only really once release behind and 10g r2 is very stable so no shame there
25 critical updates? are they for bugs which affect you, if not why fix what isnt broken?
direct io is fine - if you need it
first point sounds a bit off, but a lot of people dont use AWR.
So the real question is, is your database at risk - for that you would need to get proper external experts in to assess your business needs vs what you have
Many thanks. Sometimes it's hard to know if someone is trying to sell their services by creating an artificial need, or if they are giving you the straight scoop. Will continue to investigate. Thanks!
Both of my clients in US tend to apply CPU patches to fulfill the SOX requirements. Of course with due testing in dev, test, and qa environments before applying in production.
Thanks,
Vijay Tummala
Try hard to get what you like OR you will be forced to like what you get.
On one hand, blindly applying patches is the best way to keep awakening bugs; on the other hand different regulations force DBA to apply CPU patches - at least in the U.S.
Disclaimer: Advice is provided to the best of my knowledge but no implicit or explicit warranties are provided. Since the advisor explicitly encourages testing any and all suggestions on a test non-production environment advisor should not held liable or responsible for any actions taken based on the given advice.
There is also the fact that any bug related TAR opened with Oracle on a non-fully-patched DB is almost guaranteed to result in a patch-first-and-try-again answer.
There is also the fact that any bug related TAR opened with Oracle on a non-fully-patched DB is almost guaranteed to result in a patch-first-and-try-again answer.
yeah... and once you do that the next answer from Oracle support would be "upgrade"
Disclaimer: Advice is provided to the best of my knowledge but no implicit or explicit warranties are provided. Since the advisor explicitly encourages testing any and all suggestions on a test non-production environment advisor should not held liable or responsible for any actions taken based on the given advice.
). Obviously, with a single DBA, this isn't a huge system, but I'm not sure what metrics to give here for reference.
Reply With Quote
Originally Posted by hrishy
Bookmarks