Re: RAC + cluster waits
- From: De DBA <dedba@xxxxxxxxxx>
- To: John.Fedock@xxxxxxxxxxxx
- Date: Wed, 19 Feb 2014 14:41:33 +1000
Hi John,
I'm not very experienced with clusters, but I believe that most clusters
nowadays run on ASM. If yours does, could the wait be ASM related? As you
mention that the cluster is mature, and has been running (and presumably
growing) for several years, perhaps the ASM instance now has reached the point
where the SGA is beginning to fill up? In other words, perhaps the cause for
the waits is not in the DB, but rather in the underlying cluster filesystem?
Just my 2c..
Cheers,
Tony
On 19/02/14 00:19, Fedock, John (KAM.RIC) wrote:
Need some advice regarding cluster wait issues on a 3 node Linux 11.2.0.3
Oracle system.
Short question -- How can I determine what is causing recent cluster waits on
our database?
Long question -- We have a mature system that has been running well for several
years. Recently some performance issues have started and they seem to be
related to cluster wait events.
It is easy enough to see what events are waiting for clustered waits, but not
what is *causing* them to wait. When we are having issues, I may see several
dozen session waiting for 2 or 3 seconds and then they clear up, then a few
seconds later, it repeats. I have run several awrrpt and awrgrpt and cannot
see anything obvious.
It also seems like node1 is the one that has the session waiting.
Any advice?
Thanks all.
John Fedock
Other related posts:
