0

I inherited management of a zimbra server whose backup system is in an unknown state. Looking at the manuals zimbra provides commands to perform backups at the application level which would be great but I am lacking disk space to use them.

So before I start moving things around to properly provision the VM and potentially breaking something in the process I would like to get a failsafe copy. Can I just get a verified copy (using rsync) of the /opt/zimbra directory and will that get everything that would need to be restored in the worst case?

  • 2
    Most likely not. The Zimbra setup will rely on other things, which might be as small as one or more system users created or potentially much more complex issues. – Sven Jul 24 '17 at 21:47
  • Are you using the OpenSource version or the Network edition, because that makes a big difference in your backup strategies. – Micha Kersloot Jul 25 '17 at 8:31
  • @MichaKersloot I am not sure -- zmcontrol -v returns Release 7.2.6_GA_2926.RHEL5_64_20131203115905 CentOS5_64 FOSS edition. does that tell you? – AlanObject Jul 25 '17 at 20:39
  • Yes, that tells me it's the FOSS (Free and Open Source Software) version. – Micha Kersloot Jul 26 '17 at 9:50
  • 1
    The Zimbra Network edition backup is based on the ZeXras suite Ximlet/plugin, which includes both backup and migration tools. It appears that v. 7.2.6 in on the compatibility list, and the suite is not very expensive. Support I think is also available. – Colt Jul 26 '17 at 13:28
1

You can make a 'backup' by copying all data from /opt/zimbra. Make sure the data is not in use by shutting down Zimbra completely or create a snapshot with the correct system tools.

To use the copy on a new/different server, you need to install exactly the same version of Zimbra on that new server and then overwrite /opt/zimbra with the data from your copy.

1

Just take a snopshot of the VM before you experiment with the zimbra upgrades, or whatever you mean by "moving things around". On failure, just revert to the snapshot. If everything goes well, you have zero downtime

  • Unfortunately I don't have enough disk space in the Repository to even take a snapshot. To mitigate that means to copy the whole VM image and I wanted a fall-back in case that goes wrong. – AlanObject Jul 28 '17 at 18:33
1

You can copy completly the /opt/zimbra but need do and additional step with the ldap db. You must export it from the origin server and import it on the target server.

Another option to use a second server to do a rsync complete from all the /opt/zimbra, but the zimbra services must be stopped. I do that on servers with 200 users or less. The origin server stop all the services at 2AM, do a rsync to all the zimbra folder to the second server, and when finish start zimbra again and all work fine. At next day, i check some accounts in the destination server to check if the precious day messages are displayed correctly.

  • good idea. I don't want to stay up all night ;) but I can try a cron job to do it. – AlanObject Jul 28 '17 at 18:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.