I hate to break it to you guys, but this feature probably will never be implemented. It has been part of the official roadmap since the very start of the project. (see this document from 2005:
http://files.zimbra.com/website/docs...20Overview.pdf)
Also, the RFE has been in bugzilla since at least 2006 (
Bug 11423 – disaster recovery through server to server sync (beta)) and has amassed nearly 100 votes.
The feature just seems to be continually pushed back to some feature release. An official stance from Zimbra on this seems long overdue. I consider the fail-over solution using a SAN and RHEL cluster services to be inferior, since these produce added complexity and issues of their own.
Come on Zimbra make a choice: drop it or implement it!