I'm trying to set up RHEL Cluster. Everything is fine (well, not at all).

I'm using 2 nodes, and 1 quorum (LUN). When I start both cman services at the same time, it's fine. Cluster is up, every nodes are UP.

But if I reboot one node (or get him fenced), I still stuck at "Waiting for Quorum' and timed-out. So cman will not start.

I tried to modify votes (1 each node, 1 quorum, 3 expected). I also modified /etc/sysconfig/cman to comment out cman_quorum_timeout and set it to 60 sec.

I don't really know what I'm missing, and how to troubleshoot, since there is nothing interesting in log (messages, qdiskd, corosync...) even if I put loglevel=4 in cluster.conf.

Thanks in advance.


Ok. It was a network related issue.

RHEL Cluster is using Multicast, and Cisco switches we got we're using IGMPv1. It appears that when no QUERIER is defined, Catalyst switches let multicast discussion open for 180 sec before timed out.

As we can't put a Multicast Router, we decided to use the transport=udpu in the /etc/cluster/cluster.conf on the <cman> node. So cluster will use unicast.

No more issue now.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.