Network Replication of VM after vMotion

[mwpreston]

Hi, Hopefully I can get some help with this one...it's not a big deal, just need some clarification as to what is going on...

We have our SQL server that hosts the Virtual Centre DB virtualized. Therefore, the way that I added this vm to replicate is as follows.

1. I added the ESX host that this VM sits on individually ( the sql vm is sticky to that host )
2. I setup a job to replicate it using Network mode. It wouldn't work in Virtual Appliance mode.

During our upgrade to 4.1 we had to vmotion that sql box off of the host it is normally on, upgrade the host, and then vmotion back. After this Veeam was reporting 'Object not found' during backups/replications. No big deal really, I just re-added the vm in the job again and everything was fine...sort of... It added a new folder in my replication datastore for that vm (so essentially I have 2 objects in the replica's listed for that job) It started to replicate to the new folder and slowly delete restore points in the older one, which is fine, however, it has left one restore point in the old folder and it isn't going away as retention kicks in. I'm not sure if it is safe to delete this or not...

The reason I bring this up now is that we had to vmotion that same vm again today, and when it returned to it's original home, Veeam is now reporting 'Object not found' again. I'm not sure if there is away that I can prevent Veeam from creating a new directory again when I re add this VM, or if that is just the way it works. I would like to prevent yet another folder containing one restore point again. However, if it is safe to delete the restore point i will wait...

Make sense...let me know if you need more info...

Thanks,

[Gostev]

In order to make Veeam Backup VMotion-aware, the jobs must be configured through vCenter. However, as you know this is not possible. So unfortunately this setup right now would require re-adding VM to the job after each VMotion.

Sure, you can delete old replica folders at any time. You may also want to delete "old" VM from the job settings before adding new one, as you do not need to keep it.

Thanks!

[mwpreston]

OK sounds good...

Is there any way to remove the restore point that is hanging around from the replica...I can't seem to find it if there is. What are the consequences if I just simply delete the VeeamBackup/VM-NAME(##) folder that is hosting the old replica?

[mwpreston]

On another note, is there a way to storage vmotion or move veeams replica's to another datastore? Is it simply a matter of removing the replica from inventory, coping the files and re-adding the VM to your inventory. Or can you use svmotion capabilities to do so? Will the vrb's copy when it is storage vmotioned?

I know this doesn't fall in line with the original question, but thought I would ask it anyways...

[Vitaliy S.]

Hello Mike,

The only way to remove a restore point from you replica is either to use a retention policy, which is applied at the end of the each job run or remove VRB files manually from the datastore, though it will keep all the records about the restore point in Veeam SQL configuration database, so I would avoid doing that.

If you need to remove the previous replica, you may want to choose "Delete from disk" option on the replica record within the backup console. It would delete both... the records from the SQL configuration database and the actual files on the datastore as well.

As regards to your second question about moving Veeam replicas to another datastore, then it could be made (VRB files won't be touched during the storage VMotion process...), but everytime you re-add a vCenter Server VM to the job, new replica files will be created, because you'll have a different VM ID on each of the hosts. To take it in another way, it will be treated as new VM each time.

Here is an existing discussion of how to backup/replicate vCenter Server VM, might be useful:
http://www.veeam.com/forums/viewtopic.p ... 988#p19663

Thank you!