Re: Huge CPU Wait spikes displayed in EM12c
Date: Fri, 31 Aug 2012 10:09:00 -0500 (CDT)
Message-ID: <0556e676179d98344f59003b5a23d692.squirrel_at_society.servebeer.com>
Karl writes:
> CPU Wait is "unaccounted for DB Time" on high run queue or CPU intensive
> environments. This is similar to "unaccounted-for time" on 10046 trace
> which is a result of time spent of DB calls preempted (sitting the run
> queue wanting CPU but unable to get it).
I smell a learning opportunity here! Those waits are interesting to me because I can't find any corroborating evidence in AIX that it's an issue. The only thing I see is mild page fault spikes, but I've yet to correlate them with anything, as I haven't found that it's been overtly detrimental to performance.
Then again, from the AAS screenshots in your links, our servers are on vastly different scales. :)
Thanks much! That gives me the perfect bit of info I need to continue investigating this (in my "spare" time!).
Rich
-- http://www.freelists.org/webpage/oracle-lReceived on Fri Aug 31 2012 - 10:09:00 CDT