Backup VM Disk Layout?

[ferrus]

Is it possible to backup a VM disk config, without the actual data?

We've recently moved three VM backups from Veeam B&R, to Veeam Agent in-guest - but now I'm trying to work out the DR workflow.

I understand that Veeam Agent isn't a P2V tool, but the servers we are backing up have >25 disks each, and can change sizes several times a year. So having a blank failover may not be ideal.
I could continue the B&R backup with the App processing switched off - but that would still process over 22TB of data.

Is there any better way that the duplicate VM, for a rapid DR restore from a Veeam Agent backup - to VMware?

[HannesK]

Hello,
that's an interesting question, but I'm afraid there is nothing "out of the box" that you can use.

What I would do: write a powershell script that reads the VMware configuration and that creates a new "spare-VM" were you can restore with the agent.

Best regards,
Hannes

PS: what's the reason to move to agent based backup? In general, I see no problem doing VM backups of VMs with 22TB. We have many customers doing backups of larger VMs...

[ferrus]

Thanks - I like the idea of the script, I'll look into it.
I wonder if there's anything that could be done in VMware, by backing up the .vmx file - and creating a VM from that.

I've got another thread in the Veeam Agent for Windows forum, but we've been struggling with our VM based Exchange backups for some years.
We've optimized our storage and environment as much as we can, including all the VSS registry keys - but he still have VSS timeouts during the VMware snapshot window. This can lead to DAG failovers, transaction log truncation failures and filled log disks.

Within the first week of using the Agent backups - we've had the best, most consistent problem-free Exchange backups we've had in years. The single disk backup rate is several times greater over our 10 GbE, than our FC Direct SAN backups.

It's only been an issue with the Exchange VMs. Probably because of the 25+ disks - which extend the VM snapshot duration.
Our Exchange Admin is reticent to change this design however, as it's the recommended MS configuration for our environment.

[ferrus]

What I would do: write a powershell script that reads the VMware configuration and that creates a new "spare-VM" were you can restore with the agent.

Think I've found a solution.

A nightly Veeam 'File Copy Job' of each VM's .vmx file, from the running VM folder - to the Veeam server storage.

In a DR situation, we could register a new VM from the backed up .vmx file, boot to the Veeam Agent Recovery ISO - and restore to the fresh VM.