Best Practices for backing up Veeam repository

[Mihai Langa]

Dear Veeamers,

In our enterprise infrastructure we currently employ 200+ VBR servers (both Standard and Enterprise licenses) that are used to backup/replicate thousands of VMs throughout many locations. The Default Repository is used only to store the VBR's configuration database, while the rest of the data, generated by the jobs, is stored on other servers that have the Repository role installed.
I'd like to ask for your opinion on what are the best practices for backing up the Veeam configuration backup database (.bco files) in the Default Repository. In case of a VBR failure, in order to recover to a fully functional VBR environment do we require to back up the additional repositories or just the Default Repository?

Thanks in advance!

Mihai

[foggy]

Importing configuration backup file into a newly installed Veeam B&R is sufficient to restore the Veeam B&R instance with all its settings.

[Vitaliy S.]

And in order to protect these files you can either use file copy jobs (targeted to a DR location) or use robocopy scripts pointed to the same destination.

[liviu.tutuianu]

By the way, what happens if the repository is lost where the metadata is stored? Should Mihai also backup all those repositories to a central location so he will be able sure that nothing is being lost?

[foggy]

Are you referring to replica metadata? If it is lost, it will be re-created during the next job run.

[LarsWulf]

When it comes to the question how to backup and project your Veeam Repository and all other components Veeam Backup Endpoint Protection comes to my mind.

I think many people will be using physical server for Veeam Repository and Proxys because of direct SAN access and being separated from the infrastructure Veeam is used to protect.
Currently I personally only use the Configuration Backup function to save my config to a repository and have it on a different physical unit then the Veeam server itself. But in case of an failed server I will need to start with installing Windows etc. So recovering the backup infrastructure itself is a quite manual and time consuming process as you have no full backup of your Veeam components only a config backup.

But with Veeam Endpoint Protection you could also backup your physical Veeam servers. And in case of an failing Veeam server you can easily recover these from full backup instead of reinstallation.

Maybe Veeam could somehow integrate Endpoint Protection into Veeam Backup & Replication to easily protect your physical backup infrastructure through a single point of management.

[Vitaliy S.]

Hi Lars,

Yes, that's an interesting use case for Endpoint backup, thanks for sharing.

Thanks!

[dellock6]

I've just read this thread now, and I have a different vision...

If the need to protect the repository is for high availability and survive the loss of the repository itself, honestly any backup activity will require too much time to restore it. We are talking about a machine (virtual or physical doens't matter) that can be several TBs in size, and it will take an insane amount of time to be restored. And instant VM recovery is useless because it will get us in a catch 22 situation...
I'd prefer to replicate the repository itself either via backup copy jobs or some lower-level replication system, like scale-out architectures, storage replica, replication software, and so on. With this I can loose one of the nodes of this cluster, and still have all my repositories in place.

If I can't afford this design, then I'd go for backup copy job or even tape, because "maybe" (don't want to offend anyone) I have a low budget also because I don't have strict RTO requirements, and so the downtime to first of all restore the repository is acceptable. But again, there are nowadays effective solutions to create redundant repositories without spending a fortune, and they can also be replicated geographically to have geo-redundant solutions, good for this multi-site scenario of the OP.

[LarsWulf]

Hi Luca,

you are absolutely right in case of an DR event a backup of the repository is useless. There I need a different solution like you described. And also making the Repository data with several TB HA this solution is not the right one.

But I am just thinking about the Windows installation including the Veeam components and SQL installation.
For example in my case I have 6 physical servers running. 4 Proxy only servers, 1 Repository Server with repository data on FC attached storage and 1 MGMT Server with Veeam Server, Enterprise Manager, SQL etc. And the only thing that is backuped up of these 6 servers is the Veeam Config backup to the repository on my FC SAN attached repository. So if one of these 6 servers fails I need to reinstall Windows, join it backup to the domain, install Veeam components etc. This is of course possible but will take some time and is a quite manual process. If I would have a valid backup of the entire Windows installation I just could insert the Veeam Endpoint Recovery disk and start the restore. Really less steps to do and so much faster.

So the idea is to protect your Veeam Services not the repository data.

Regards
Lars

[Vitaliy S.]

If the need to protect the repository is for high availability and survive the loss of the repository itself, honestly any backup activity will require too much time to restore it. We are talking about a machine (virtual or physical doens't matter) that can be several TBs in size, and it will take an insane amount of time to be restored. And instant VM recovery is useless because it will get us in a catch 22 situation...

The OP was talking about of how to protect our config data, which is small and should not take much time to backup

[Mihai Langa]

Dear All,

Thank you very much for your assistance and thorough replies! Having only to copy the VB&R server configuration data to a DR location really simplifies things for us.

Best Regards,

Mihai