Any idea why? Don’t give simple suggestion “Just upgrade to latest OS”. We have application there not stable on new version of SLES. Currently we use SLES 11 SP2.
offline/unclean usually is caused by interrupted communications between the nodes. Have you verified the communications link is up & running as expected? How’s the state of the ring(s)? Anything in the logs that might point to the root cause?
If everything else fails, monitor traffic for the configured (multicast?) address on both nodes, to see if the other node’s packets get through.