vMotion VM in Veeam DataLab

[kapple]

Hi,

I understand that you can do a staged restore, which utilizes the Veeam DataLab (formerly VirtualLab).
I am wondering if it is possible to vMotion a VM that is running in a Veeam Datalab into production.

The reason I am asking is that I have been running Windows Updates on a VM that is woefully out of date, within a Veeam DataLab (Launched as a SureBackup job). After that, I want to upgrade the OS.
I am most of the way through this process, and now I realize that to use the Staged Restore capabilities of Veeam, I need to initiate a separate restore job. I had thought that after launching a SureBackup job, I could use a staged restore after I made my changes (I need to thoroughly read up on this stuff before I start working on things I have never done before).

I could start over, and just use a "Hello World" script so I can start the staged restore job, do the updates and upgrades I want to do while the DataLab is running, then finish the restore.
This is how it works, correct? The backup will launch in a DataLab, I will be able to make changes to the restored VM, and then finish up the restore job?

However, since I have a SureBackup job that has been running for a couple of days while I work on upgrading the VM, I am hoping there is a method of restoring the VM in the DataLab into production.
Please let me know if this is a valid path forward, or if I need to start from scratch.

Thanks,

[Egor Yakovlev]

Hi Ken.

Yes, you can simply vMotion said SureBackup'ed VM from Veeam DataLab to production datastore and reconfigure it's vNIC to map it over to respectful production network.
However, also note that your production "original" VM is still there, I guess? Was it working for last 2 days while you was playing with DataLabs? If so, those 2 days of changes in production are not reflected in the DataLab machine copy, thus your idea to migrate from 2 days old DataLab might lead to some data loss.
After migration, hit Stop Session in VBR's SureBackup Job window, and it will dismount leftover datastore that we used for SureBackup launch from the environment.

/Cheers!

[kapple]

Thank you Egor,

There is no changed data that would be of importance, this server just runs an application, and the database is hosted on another server.

That said, is there any reason to start over with a staged restore? I am assuming that vMotioning a VM out of the DataLab will result with a VM that would be the same as a staged restore, with a changed name which I could edit once it is back in the production environment.

Was I correct in my assumption of the staged restore, that the VM will come up in the DataLab that was configured in the job, I could then make the changes to the OS requiring several reboots, and then finish the restore?

I also realize that you must include a script in the staged restore job, so a simple "hello world" script should meet that requirement, correct?

[Egor Yakovlev]

You are correct everywhere, except one part with Staged Restore:
- "is there any reason to start over with a staged restore?" - no reason! keep your already done work and use vMotion.
- "I could then make the changes to the OS requiring several reboots" - no, Staged Restore script must run uninterrupted, thus a Staged VM reboot will break it's execution. Staged Restore is meant for automation cases where you need to, for example, delete sensitive data prior to production recovery, or add new antivirus signatures to the server. Windows Updates are not the best use case for Staged Restore, and are usually covered with other tools, like Microsoft NAP(Network Access Protection) - say you restore outdated VM with Veeam, NAP detects missing updates, automatically puts it to quarantine network with WSUS, and after required updates were pushed to the machine, it puts it back to production network.
/Thanks!

[kapple]

Thanks again Egor,

You gave me a much better understanding of the staged restore use case, and confirmed I can use a SureBackup job to promote a VM in the DataLab to production. You have just made life much easier for me!

[kapple]

Hi, I just want to give another update.

It appears that you cannot vMotion (host/compute resources, storage, storage and compute) a VM out of a Veeam DataLab. When performing a storage vMotion, there is a warning stating "Snapshot directory settings will be reset to the VM's home directory after the migration." When vMotioning to another host, there is an issue of the NIC used within the DataLab not being available in the production environment.
The storage vMotion finishes within a few seconds as reported in the vSphere client, which is way to fast for this to actually happen. I have tried this with the VM both powered on, and shut down.
I have had a couple of different outcomes when trying to vMotion a VM out of the DataLab. One was that the lab just shut down. The other is that the lab just kept running. In both instances, there was no vMotion of the VM into production.

That said, I can, and did clone a VM out of a DataLab. Is this the supported method of migrating a VM out of a DataLab into the production vSphere environment?

Thanks,