Hyper-V: Instand recovery and work with VM

[omfk]

Hi all,

regarding
https://helpcenter.veeam.com/docs/backu ... ml?ver=110
I have a small question:
If I recover a VM with this approach, will I'll be able to work with that VM; i.e. can I delete/add/modify data or do I have to fully restore the VM to a different location for this? The VM in question is mostly a file server.

BR
Frank

[Mildur]

Hi Frank

The VM is fully available for you to work with it as the original one was. If you have choosed Restore to the original location, you can access it with the same hostname and ip.
The only difference to the original vm you have, the requested data needs to be loaded from the restore point when you access it. The read performance will not be the same as the original vm.

You must finalize the instant vm recovery to move it entirely back to production storage.
Finalize Instant Vm recovery

[omfk]

Hi Mildur,

thanks for the reply.
Just to be sure on my side: With this approach I don't need any free space (or only very little) on the backup server itself? In our case for a 10TB file server I don't need 10TB free space for instant recovery?
Of yourse I'll need this later as I consider this option only for temprary use until the replacement server is on side or the original server repaired.

BR
Frank

[Mildur]

Hi Frank
Your welcome.

You don't need space on the backup server to start the instant vm recovery. Only a little bit for the few logs of course.
But for your fileserver, you would need the 10TB on your hyperv storage to start an instant vm recovery session.

At the start of the instant vm recovery, Veeam will preallocates the required disk space for the vm on the production datastore of your hyperv cluster. If you choose Restore to the original location, the old virtual disks will be deleted and you can reuse the 10TB storage.

If you choose Restore to a new location and not have deleted the old fileserver, then additional 10TB will be required for the instant vm recovery.

[omfk]

Hi Mildur,

I'm afraid I'm a little bit lost here:-(
The setup:
1 production server with several VMs, among them 1 big file server
1 Backup server which shall hold Veeam and the backup files
Production server goes down.

I was now under the impression that I can "restart the VMs on the backup server" itself (so Veeam magically works with and within the vbk/vib files) until I have a replacement. So on the backup server I only need a few Gigs for logs (let it be 100GB) for this process.
If I later want to restore the VM on the replacement server of course I need real space.

Is my impression wrong?

BR
Frank

[Mildur]

Hi Frank

Instant VM Recovery needs a working HyperV Server to start the vm.
Ideally this is your productive hyper v server.
If you have lost your hyperv server, you need to deploy a new one, before you can use instant vm recovery.

Veeam cannot start a hyperv VM by itself without the hyperv components.

[omfk]

Hi Mildur,

that's understandable.
The backup server itself is a physical server running Hyper-V with Veeam as one VM and a small Windows VM which acts as secondary domain controller.

Will that change anything on your answer?
I don't want to keep 20TB of free space on the backup server in order to restore the file server on it. I was hoping to cover the time until repair/replacement to bridge with my impression of instant recovery.

BR
Frank

[Mildur]

You can use this hyperV with the Veeam VM for doing an instant vm recovery of the filer.

But you will need to have 10TB storage and some storage for the checkpoint (changes on the server during the instant session) on the Target Datastore on the hyperv server for the VM at the start of the instant recovery session.

Choose Target Datastore

[omfk]

Hi Mildur,

thanks for the clarification.

So the advantage of instant recovery vs. "normal restore" lies with the fact that I already can work with the VM while restoring and I don't have to wait until the end?

Not what I hoped for, but of course still a great improvement.

During instant recovery can I choose only to "restore" certain parts of a VM. With my fileserver I would then be able only to restore the important part and not all historic project data. Or would that require different backup jobs?

Backup job 1: full VM with recent data
Backup job 2: only historic data (here it would be sufficient to make it once per week)

BR
Frank

[Mildur]

So the advantage of instant recovery vs. "normal restore" lies with the fact that I already can work with the VM while restoring and I don't have to wait until the end?

Correct, that‘s one of the main purposes of instant vm recovery. You can start a vm from the backup file without waiting until every block is restored.

During instant recovery can I choose only to "restore" certain parts of a VM.

Instant vm recovery will always start the entire vm how it was protected by the backup job. As far as I know, virtual disks can not be excluded. I can‘t confirm it right now in my lab, but I‘m pretty sure it is not possible.

On the file and folder level it‘s definitely not possible. The purpose behind instant vm recovery is to provide access to the entire vm when a normal restore is taking to long.

Backup job 1: full VM with recent data
Backup job 2: only historic data (here it would be sufficient to make it once per week)

HyperV VM Backup Jobs will always backup the entire virtual disk. You cannot choose to only include a part of the data on a single virtual disk.
If you have dedicated virtual disks for most recent data, and a second virtual disk for archived projects, you could create a second backup job to backup the archive vdisk and another job for protecting the os virtual disk and all other vdisks.

[omfk]

Hi Mildur,

a big thanks for the clarification.
Regarding the exclusions for the backup job I was misdirected by the options of the Windows agent. Different virtual disks is something to look into, but probably a project for next year (as this require some brain work on a different programm - not Veeam related).

BR
Frank