[Linux-cluster] Two-node cluster unpatched B doesn't see patched A

[Lon Hohberger]

On Thu, 2008-02-07 at 17:08 -0500, Sutton, Harry (MSE) wrote:

> On my production cluster, which uses a Quorum Disk in place of the third 
> node, I completed steps 1 and 2 on Node A, but the cluster did NOT 
> reform. cman sends out its advertisement, and I can see that Node B 
> receives it (by looking at the tcpdump traces), but Node B never responds.

That's strange, I hardly think it's related to qdiskd.

Maybe check 'cman_tool version' on both and ensure that the cluster.conf
version matches the output of 'cman_tool version' - that's a common way
to prevent the cluster from forming on RHEL4.


If they're out of sync:

  * get the right cluster.conf in place & bump the config version #
  * run 'cman_tool version -r <new_config_version_#>'
  * copy cluster.conf to other node

Note: Restoring the nodes to the same package versions will have utterly
no effect if the config version is out of sync!

-- Lon