zigzagdna_at_yahoo.com wrote:
> No Oracle backup jobs or database statistics run at that time. SAN is
> used for disks. I have to check details of SAN with storage group. I
> will look into Filemon, but I have been using Windows perfmon which
> lets me collect disk usage, memory usgae, cpu usage etc
> periodically .It is pretty nice utility.
>
So it is shared storage, I'd definitely be looking at a scheduled task
somewhere in the infrastructure. It does strike me as odd though,
especially if you say you've asked users to hold off between 3 and 4
that you are doing far more i/o at that time than you are just before. I
wonder if it turns out that there really is nothing else sharing
components of the san that affect these boxes, if the larger io
throughput is responsible for exhausting san cache. I'd still wager on a
scheduled task somewhere in yur infrastructure though.
--
Niall Litchfield
Oracle DBA
http://www.orawin.info/services
Received on Thu Oct 25 2007 - 16:05:04 CDT
