HyperV failover with Veeam

[geets]

Hello,
we have HyperV failover configured for our VMs.
So for every VM called xxx theres a VM called xxx_replica on one of our other HV-hosts.
This is all HyperV-native.

This produced some difficulties with veeam since we didn't want to backup the _replicas. We managed to use power shell to dynamically add the replicas to a vmgroup. We then excluded that vmgroup from backup jobs in veeam backup and replication.

This is getting ugly with copy jobs and long time retention.

Is there a better way to handle failover/replicas in Veeam?

Thanks in advance
Alex

[PetrM]

Hi Alex,

I think that the best way to handle it is to run both replication and backup jobs from the single application:)

Please consider the following options:
- Replication from backup: backup files will be used as a source of data for replication instead of reading the same data twice from production environment
- Flexible failover and failback: plan entire failover in advance for VMs running interdependent applications or perform maintenance on your production hosts with planned failover
- Easy to manage: you can add several VMs into the single replication job which has its own schedule and other settings

And regarding the following issue with the current setup:

This is getting ugly with copy jobs and long time retention.

Did you try to select VMs from backups and not from infrastructure when you add objects into the backup copy job?

Thanks!

[geets]

Hi, thanks for your reply.
I understand that using only Veeam B&R for both would be a clean solution. Unfortunately this is not possible due to administrative limitations. We have to keep failover and backup separated. Failover will be managed by HyperV and Backups by Veeam.

Perhaps you can shed some light on something where I have concerns:
- I have a VM called vm01
- It runs actively on HyperV-A and has its failover-replica on HyperV-B
- Veeam is configured in a way, that does NOT backup failover-vms. - Thats what we want.
- When a failover happens to HyperV-B, Veeam will start backing up the now active VM on host HyperV-B. - Also how we want it.
- Important: The target repositorys for both of those VMs (active and failover) are different since they come from different HyperV-Hosts:
Before failover they go to HostX:\\HyperV-A\vm01.
after failover they go to HostX:\\HyperV-B\vm01 - Is here a problem already from a best practices point of view?

- Now we do Copy Jobs with GFS. Will those Copy Jobs copy the backups in the old place AND the ones in the new place? Is Veeam aware that they are not different VMs but just 2 different days of backups of the same VM?

To your Question:

Did you try to select VMs from backups and not from infrastructure when you add objects into the backup copy job?

We have "from jobs" configured right now.

Thanks in advance!

[PetrM]

Hi Alex,

Is here a problem already from a best practices point of view?

As far as I understand different backup jobs process different VMs: original or active replicas. I wouldn't expect any issues but I never tested the scenario with the native HyperV failover.

Will those Copy Jobs copy the backups in the old place AND the ones in the new place?

Backup copy job does not know about failover, replicas or other infrastructure specifics, it copies data from a primary repository to a secondary one.
It always copies data to the specified target.

Unfortunately this is not possible due to administrative limitations.

Would you be so kind to clarify what exact limitations do you have and why do you separate backup and DR strategies?

Thanks!

[geets]

Hi, thanks for your reply

Would you be so kind to clarify what exact limitations do you have and why do you separate backup and DR strategies?

maybe I used the word "administrative" in a wrong way... I'm no native speaker.

There's two issues when it comes to consolidating both backup and failover within Veeam:
One: theres two teams involved, the Hypervisor team and the backup team. It would mean we'd have to reorganize things from a responsibilities point of view.
The other reason for not using solely Veeam is more of a technical one: besides gfs and copy-jobs, the system works fine as it is. It'd probably be a lot of work to migrate our management of VMs and management of replication (failover) from HyperV to Veeam.
The 'migration path' is very unclear to us and I have concerns, that we'd have to reconfigure every backupjob by hand and in the end miss out on some old backups and so on.

[foggy]

You could just create new replication and backup jobs and leave existing ones disabled and intact for some time until the new jobs will reach the required number of restore points Thus you will ensure you don't miss anything.