HP insight manager daemon cmaidad causing Kernel crash

[Yohann Dumais]

Currently investigating an issue where we had a server running Oracle on a RHEL 3 crashing during backup process.

I noticed that the cmaidad process was using 15% cpu just before the kernel crash.

Investigated and the DB Admin did mentioned he started using the HP insight management tools more this last month.

To top it off, these are a Oracle rac server, db partition are on ocfs, possible conflict with cmaidad?

Anybody experienced such problem combined with heavy io usage?

[dirk dierickx]

would be surprised if cmaidad caused a panic, probably more the fault of ocfs. are you on the latest patches for RH3? ocfs os constantly being updated certainly at the time RH3 was released.

[Yohann Dumais]

We have the latest ocfs version 1.0.14-1 (we don't consider moving to 2.x).

We got the usual response from RedHat and Oracle to move up to latest release. We would have like to identify the problem instead of just upgrading and waiting for another occurrence.

Current RedHat 2.4.21-27.ELsmp, we will upgrade to 2.4.21-47.0.1.ELsmp.

Current Oracle 9.2.0.4, would update to 9.2.0.7, this one is more inconvenient.