[SOLVED] Cloud And Disaster Recovery

[crunch]

Hello! Been running Zimbra for 2 years now inside a VMWare Server 2 instance for 25 users. Works fabulous! Although it's hard to convince users that the power of Zimbra lies in the Zimbra Web/Desktop client

I need some help regarding backing up to the cloud...

I'm working out a backup/recovery plan that involves using Jungle Disk (rackspace) to capture daily changes. A full Zimbra backup is about 14GB and it would take 2-3 days to upload all of that and then do it again in 2 days. That doesn't seem like a good way to do it. I'm already capturing local backups and just want to have a full disaster (loss of equipment or entire building) plan in place.

I have 2 questions...

1) If I keep a snapshot of /opt/zimbra (backups are elsewhere) will that allow me to install Zimbra on a new server (same version) and just restore that directory on top to get back up and running? Would that require Zimbra to be shut down before each backup?

2) The full backups happen every 5 days. Can I do an rsync between a "master" full backup and subsequent full backups? The idea being I only want to upload 14GB once and then capture changes to the full backups that Zimbra produces.

If you need further clarification please let me know. All input welcome.

Thanks!

Chris

[crunch]

I think I have it licked unless someone has some caveats? I managed to try a /opt/zimbra replacement by installing the same version and replacing it with a backup copy. After some messing around with a few issues it seemed to work like a charm. Didn't do a complete sending/receiving test but all the accounts showed up.

So now all I need to do is configure Jungle Disk to backup my backup of the /opt/zimbra directory and it should be good to go. 24 hours of email loss is acceptable unless I get ambitious and start syncing redo logs.