Re: RE: Sun Cluster - Oracle: aiowait timed out

From: techinfo@qatar.net.qa
Date: Sun Jan 27 2008 - 02:27:12 EST


Dear Sun managers,

Many thanks for responding to my query and providing valuable inputs.

I just found that one power supply has gone faulty, and cache has been moved from 'write-behind' to 'write-through' state. Changing the power supply has made the DB performance better though the aiowait time error was appearing even before we have DB performance issue.

But I have a query:

SE6320 array has redundant power supply. Should one power supply going faulty change the cache status for that array to 'write-behind' disabled? Going one power supply faulty will always have performance impacts? Any fixes or work-around for this without compromising on data integrity? Any inputs from storage experts will be highly appreciated.

Regards,

./Ansu

----- Original Message -----
From: "Engle, Victor" <Victor.Engle@netapp.com>
Date: Thursday, January 24, 2008 4:44 pm
Subject: RE: Sun Cluster - Oracle: aiowait timed out

>
> Just a guess, async I/O timeout. Have you looked in
> /var/adm/messages to
> see if anything was being reported with regard to fibre channel or
> scsierrors. If possible, collect some vxstat info during the next
> event.Perhaps it is purely load related and may be mitigated by
> changing the
> storage configuration. Of course Oracle is especially sensitive to
> performance of storage that the logs sit on.
>
> Good luck and I look forward to your summary.
>
> Regards,
> Vic
>
>
> > -----Original Message-----
> > From: techinfo@qatar.net.qa [techinfo@qatar.net.qa]
> > Sent: Thursday, January 24, 2008 6:55 AM
> > To: sunmanagers@sunmanagers.org
> > Subject: Sun Cluster - Oracle: aiowait timed out
> >
> > Dear Sun Managers,
> >
> > We are facing a problem with our Sun setup.
> >
> > We are running Oracle 9i on Sun Cluster 3.1 with VxVM. Shared
> > storage is on the Sun 6320 array.
> >
> > We are facing some performance issue with Oracle
> > intermittently. Whenever we have the performance issue, we
> > are able to see the following error in the alert_RADH.log
> > under Location: $ORACLE_BASE\admin\RADH\bdump
> >
> > Error message: WARNING: aiowait timed out
> >
> > We want to know what exactly this error means, and what to be
> > done to fix the slow performance problem.
> >
> > When we checked with Oracle, what they are saying is that it
> > is an operating system. I posting this query here hoping that
> > some will be able to help me out.
> >
> > Any help would be highly appredciated
> > _______________________________________________
> > sunmanagers mailing list
> > sunmanagers@sunmanagers.org
> > http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers



This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:42:43 EDT