Backup best practice for single Hyper-V Host

[spug101]

Hi.

I know that this has probably been asked a billion times, but I’m just looking for some clarity for best practise on backing up a single Hyper-V host and its VMs.

My current setup is a single Hyper-V host with a variety of VMs (DC, Storage Server, VDIs, VBR). One of the VMs on the host is the dedicated VBR Server. Currently my setup is as follows:

1 – Primary backup job backs up all virtual machines on the host to external USB drive (including the VBR Server) (I know that this is clearly best practise for granular restore in non-disaster recovery scenarios)
2 – Windows agent backup of the host (in this case I do a file level backup and include all drives and exclude the folders holding the VHDs & VMCs (as I clearly don’t want to duplicate the data))
3 – Create recovery media for the physical host for disaster recovery (and store off-site)

The primary issue with this setup is for disaster recovery scenarios; yes, I can restore the hyper V host using the recovery media, but how do I then restore the virtual machines considering the VBR server is one of the virtual machines themselves? Clearly, I don’t want to duplicate data by backing up all virtual machines whilst also backing up the host including all virtual machine data, but the way I do it, it seems to be a chicken/egg scenario!

Is my setup just wrong? Please provide your thoughts on best practise without duplicating backup data. Thanks.

[Mildur]

Hi Frank

1 – Primary backup job backs up all virtual machines on the host to external USB drive (including the VBR Server) (I know that this is clearly best practise for granular restore in non-disaster recovery scenarios)

You should have at least two copies of your backups, with one being immutable or air-gapped. External USB drives are not the most reliable option to store backups, because in theory a lot can happen to them and they contain cheap disks, especially consumer market-focused USB drives. Do you have a second copy of your backups?

Best practice for the backup server is to use our configuration backup and restore it to a newly deployed backup server for DR. You can protect the backup server with a VM job, but we recommend against it. If you still decide to do so, do it outside of the backup window. You can find more information in our KB2645 and help center.

2 – Windows agent backup of the host (in this case I do a file level backup and include all drives and exclude the folders holding the VHDs & VMCs (as I clearly don’t want to duplicate the data))

Do you really need a backup of a single standalone host? The host doesn't have any production workload running on it besides the virtual machines. I would recommend just deploying a new HyperV host and starting to restore your virtual machines. With the right documentation, deploying a new HyperV host should not take long. And if you have a low Recovery Time Objective (RTO), such as 1-2 hours, I suggest looking at HyperV Clusters configurations or Veeam Replica to a second HyperV host.

The primary issue with this setup is for disaster recovery scenarios; yes, I can restore the hyper V host using the recovery media, but how do I then restore the virtual machines considering the VBR server is one of the virtual machines themselves? Clearly, I don’t want to duplicate data by backing up all virtual machines whilst also backing up the host including all virtual machine data, but the way I do it, it seems to be a chicken/egg scenario!

You must restore your VM backups from a Veeam Backup & Replication backup server. Simply install Veeam Backup & Replication on any machine in your environment, import your backups from the external USB disk, and restore the VMs.

My colleague Hannes wrote a blog about it: https://www.veeam.com/blog/restoring-in ... veeam.html

Best,
Fabian