I really hope this is getting fixed soon because it breaks the Cluster capabilities...
Node 1 was running and fails (of any reason you whish, such as "manual testing" 8)).
Node 2 tries to start but amavisd/clamd is too slow and zmclustat considers Zimbra is not running. So it stops Zimbra on node 2. And restarts it right away because node 1 is always down. And same problem again and again.
Actually, node 1 had the same issue but fixed itself (don't as me how, I don't know). We tried to copy the databases from node 1 to node 2 but it did not fix the issue on node 2.
Does the timeout changes also works for RHCS setup ?
Oh, and BTW, having a single ClamAV database (ie putting it on the shared SAN) could be a good idea. Right now, you can have different ClamAV databases on the different nodes... Not good.
From what we know about it, the changes do fix rhcs.
It does :D
Just upgraded the cluster (first time), upgrade went flawlessly and node 2 now take over the cluster without any problem.
Just ran into the stats issue (4.5.5 -> 4.5.7).
After 6 months my Zimbra server exhibited the ClamAV start up problems. Again, I can confirm the fix discussed in this thread solved the problem :-)