Re: CPU Wait troubleshoot

From: Chris Stephens <cstephens16_at_gmail.com>
Date: Tue, 12 Apr 2016 14:17:54 -0500
Message-ID: <CAEFL0sxvLW28DTgwgk=v4MYJLn46tHHW8SWf5HZ+sAq=U986SQ_at_mail.gmail.com>



Where are you getting the RWP recommendations from?

On Tue, Apr 12, 2016 at 12:40 PM, Stefan Koehler <contact_at_soocs.de> wrote:

> Hi Jeff,
> unfortunately the graphic is very small, so my reply is based on what i
> can see and what you described.
>
> At first CPU Wait is the OS CPU run queue time. For more details you can
> check out Craig's blog post with an experiment about it:
>
> http://blog.orapub.com/20150210/What-Is-Light-Green-Oracle-Database-CPU-Wait-Time.html
>
> … so basically you have to look at the OS CPU run queue, if you want to
> find out the root cause. Secondly it seems like you have overloaded your
> CPUs.
> 400 (possible) active sessions (on CPU) would need round about 40 cores
> based on the Oracle RWP guideline of max 10 sessions per core (for a well
> behaved applications). I can see that your CPU core is very low, but can
> not figure out the exact value as graphic is so small.
>
> Checking the OS CPU run queue details depends on your OS.
>
> Best Regards
> Stefan Koehler
>
> Freelance Oracle performance consultant and researcher
> Homepage: http://www.soocs.de
> Twitter: _at_OracleSK
>
> > Jeff Chirco <backseatdba_at_gmail.com> hat am 12. April 2016 um 19:16
> geschrieben:
> >
> > Looking for some help on how to troubleshoot this spike in CPU Wait.
> Every so often it will see this large spike and I am not sure how to find
> out
> > what it is related to or why it is happening. It spike up to 400 active
> sessions when normally we only around 2 - 5. When I click on the graph to
> > drill in it only show the CPU usage and nothing near 400 active
> sessions. What else do I need to look out to figure this out. Thanks in
> advanced.
> >
> > [Inline image 1]
> >
> > Jeff
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>

--
http://www.freelists.org/webpage/oracle-l
Received on Tue Apr 12 2016 - 21:17:54 CEST

Original text of this message