One node experiencing heavy load(OS) in 2 node RAC

From: <kybyrd_at_mail.med.upenn.edu>
Date: 19 Jul 2005 08:06:37 -0700
Message-ID: <1121785597.342924.41430@g44g2000cwa.googlegroups.com>

I have been unable to resolve a stubborn, unpredictable problem, with one node
in a 2 node RAC system(9.2.0.6) experiencing unexplained heavy loads.

The problem surfaced after this system ran with no problems for 3 months.

The "hanging" node will prevent further session creation although existing
OS processes and oracle sessions continue.

My other node works nornally so databsae access is uninterrupted.

I am running Red Hat AS 3 and I have experienced the problem with kernels
2.4.21-27 and 2.4.21-32.

The servers are two Dell 1850s.

We are also using an Network Appliance filer NAS.

The problem seems to only occur in the hours from 1am-approx 6am. There are no cron jobs running at that time that would cause the hangs.

Nothing that would suggest a cause has been seen in the OS's messages file or Oracle logs file. Producing an RDA snapshots does not show anything
of interest.

I have an open TAR with Oracle about this problem with no resolution from
them yet.

If anyone else has seen this behavior I would like to know what your determination of the cause was.

Lastly this is my first experience with Oracle RAC so there may be a configuration item I missed. However since my configuration ran perfectly for 3 months makes me suspect that that is not the case. Received on Tue Jul 19 2005 - 10:06:37 CDT