Best practises for backing up in a Hyper-V cluster?

[Mr Dudeman]

Hi All,

I am looking for advice on how best to configure VEEAM in a Hyper-V Cluster environment.

I operate multiple Hyper-V clusters, each consisting of about 10 hosts, and 4 Cluster Shared Volumes (CSV).

VEEAM was not working particularly well in my environment, hence a call to Tier 3 support (in Europe), but now seems to be running ok.

Advice that I received from Tier 3 support is to be mindful of how many snapshots I am creating on any given CSV at any one time. Ideally I would like to keep snapshots to only 1 or 2 per CSV, but how can I configure backup jobs towards this end? Whilst I am in a position to put the VHDs onto any given CSV and arrange jobs that way this approach does not take into consideration situations where storage for a VM needs to be migrated for operational purposes.

Are there any whitepapers available with best practises for Hyper-V clusters outlined? Does anybody have any experience in this area that they would be happy to impart?

Kind Regards,

Davo

[foggy]

Hi David, doesn't the max concurrent snapshots setting meet your goal?

[Mike Resseler]

David,

I am assuming you are using software VSS and don't have the possibility to have Hardware VSS? (Or are you already using 2016 Hyper-V hosts...)

Mike

[nmdange]

So if I occasionally get VMs that fail with the error "Failed to add volume to the VSS snapshot set. Another shadow copy creation is already in progress", would lowering the max concurrent snapshots help prevent that from happening?

Also I understand in the case of Hyper-V 2012 R2 this would lower the number of VMs that can be backed up concurrently from the same volume. But since Hyper-V 2016 no longer requires snapshotting the CSV, I assume it would not have the same effect on limiting concurrency? The documentation says

For VMs hosted on Microsoft Hyper-V Server 2016, the Max snapshots value defines the maximum number of checkpoints that may be retained for a processed VM. The setting is specified at the level of a volume on which VM disks reside.

I gather that only means a single VM cannot be backed up by more than one job at a time?

[foggy]

So if I occasionally get VMs that fail with the error "Failed to add volume to the VSS snapshot set. Another shadow copy creation is already in progress", would lowering the max concurrent snapshots help prevent that from happening?

No, this settings controls the number of snapshots that can exist at the same time, while what you see I guess is just another backup running in parallel.

Also I understand in the case of Hyper-V 2012 R2 this would lower the number of VMs that can be backed up concurrently from the same volume.

No, the number of VMs backed up from a single snapshot is configured with the help of the registry values.

[nmdange]

No, this settings controls the number of snapshots that can exist at the same time, while what you see I guess is just another backup running in parallel.

I changed the value from the default of 4 to 1 and have seen this error significantly less frequently. I think Veeam is not waiting for the first snapshot creation to finish before starting a concurrent snapshot. I suspect it only happens when VSS does not behave correctly

No, the number of VMs backed up from a single snapshot is configured with the help of the registry values.

Right by default up to 4 VMs can be backed up by a single snapshot with software VSS. So if you can have 4 VMs per snapshot and 4 concurrent snapshots, then you could be backing up 16 VMs at the same time theoretically?

[foggy]

Theoretically, yes.