We are running a QA Oracle RAC on Red hat in two Dell 2950s. One the the clusters randomly reboots. In checking the messages file, I get:

Jan 8 20:47:00 quebecC2 logger: Oracle CSSD failure 134.
Jan 8 20:47:00 quebecC2 logger: Oracle CRS failure. Rebooting for cluster integrity.

The system comes backup fine and there is no interruption of service. But this is a bit disquieting. We are also running a production version with the same configuration of RAC and there has not been any trouble.

I called Dell just to see if there was a hardware issue and they said everything looked fine.

Any information regarding this problem would greatly be appreciated.

Thanks,

Mike

CRS reboots have a number of different causes. MetaLink Note #265769.1 contains lots of information on how to troubleshoot CRS reboot.

Check both systems whether they have the same I/O scheduler and the same value for a CRS parameter like O2CB_HEARTBEAT_THRESHOLD.

Always use "deadline" algorithm and set higher values for heartbeat threshold.