I am tasked with building a Zimbra NE hosting platform and would like to have as much HA and DR features as possible. All of the servers will be VMs in a vShere ESXi 4.1 cluster. I was wondering if anyone has had any success with VM Fault Tolerance on Zimbra VMs. I am concerned with Zimbra's CPU requirements and also the ability of the FT secondary to remain in sync.

I have been working with ESX for a few years and am comfortable with HA/DRS, etc. I have not yet had the opportunity to implement FT and see that this may be a good fit. Any gotchas to look for? Recommendations? I do have dedicated 1G NIC configured for FT, separate from the vkernel network for vMotion, etc.

I will be starting off small, less than 100 mailboxes, but see a possibility of this quickly expanding to several thousands.

Any and all comments, suggestions, flames welcome.