Just a quick note to say, I'm pleased I got high availability Zimbra goodness happening on two Debian virtual servers (running on VMware)

It's two Zimbra instances on an /opt partition that is synced between each other with DRBD. Heartbeat used for auto failover (if node A goes down, node B mounts its /opt partition with DRBD and takes over the virtual IP that is used for the 'main' zimbra domain that we point staff to)

Howto is on my main blog: here

Good fun :)