Restoring vCenter with Veeam’s Instant Recovery

I’ve been a long time user of Veeam, and something that sold me on it almost immediately was its Instant Recovery feature. For those of you who are unaware, Instant Recovery allows you to mount a VM from a backup (while writing changes to a delta file, thus leaving the backups ‘pristine’) without the need to copy the full VM back into production. This means that you’ll get a pretty quick turnaround on recovery – think minutes instead of an hour or more for larger VMs.

I was recently mucking about with one of my vCenter servers and I ended up pooching the SSL certificates on the server. Rather than spending the time on trying to undo the error, I opted to do an Instant Recovery restore to get things back up and running. I thought I would document the process as there are a few additional steps involved when working with vCenter.

Typically with Veeam you can just select Instant VM Recovery and all the default options will load the VM up to where the original was located. This is a problem if you are trying to restore your vCenter host as Veeam tries to connect back through vCenter to find the specific ESXi host to restore to. So, the first step is to add your destination host as a new Managed Server. You can do this by going to Backup Infrastructure –> Add Server. When this wizard pops up you’ll need to enter either the DNS name or IP address of the server you want to add. Be sure to enter a different identifier than what is in vCenter; for example, if vCenter sees that host as 192.168.1.10 then you’ll need to use the DNS name (e.g. esx01.localdomain.local).

Also note that this does not impact your licensing as Veeam is able to recognize that the “new” server is already licensed.

Next start up the Instant Recovery wizard. Once you select your VM and restore point, be sure to select ‘Restore to a new location …’. If you don’t Veeam will try and restore to the host via vCenter and will fail as per the previous paragraph. Finish off the wizard and let the VM power up. Once that is done the machine should be running and hopefully your vCenter will be back up and running. A couple of causes for failure that I have seen in the past include if the backup file is in use (e.g. if you have a Backup Copy job running), or if the original VM is still there and you have not changed the name of the VM you are recovering.

Now that the VM is back up and running, we’ll want to migrate it back into production. As mentioned above, the VM is running from within the backup file, so that file will be locked. If the file is locked then you won’t be able to use the file for the next scheduled backup, so we need to migrate the VM. You can use Storage vMotion if you are licensed for it, but if you aren’t (or can’t for some reason), Veeam has you covered.

When running an Instant Restore a new option is available in the Backup & Replication window – conveniently titled Instant Restore.

Go into this section, right click on the running VM, select the “Migrate to Production…” option (I love that the menu item is so clearly labeled), and go through the Wizard to select where you want to migrate to. Veeam will then kickoff a snapshot, copy the data to the new datastore from the backup file, and finish off with a snapshot consolidation. The VM may freeze on the snapshot consolidation mode, which is a drag when compared to Storage vMotion, but I would gladly take a small delay when you consider how quickly a VM can be restored via Instant Recovery.