As mentioned, there have been earlier developments with Heartbeat & DRBD on SME.
SME 8.0 will be based on CentOS 5, which has Heartbeat & DRBD support natively.
I've actually been looking into this myself for quite some time, and recently found a thread on the CentOS forums, which explains a failover cluster setup, which boils down to the following:
- Take 2 identicale machines (preferably, for performance' sake), or at least 2 machines with equal size HDD's & RAM and install a basic CentOS (no X & additional software). If you want to use RAID, do this as part of the installation.
- Install, configure & test DRBD & Heartbeat, making sure each machine can take over if either fails.
- Install Vmware / XEN on both machines, expanding the virtual machine to the physical machine's full size, meaning you'll be using the VM actively
- Install all the other software (webserver, mailserver, MySQL, X, etc) on top of the VM's
- let the 2 physical machines syn each other & offer high availability
Technically speaking, if you install SME 7.3 on the virtual machines, you'd have 2 highly available SME servers, since the sync will happen on the physical machine's storage. Theoretically this will work quite well, I've just not had the time yet to test it out