V10 - Backing up Proxies itself while replicating

[thomas_brun]

Hi there

We had a huge problem while backing up the proxies, at least on the disaster recovery site.

Our Infrastructure:
We have two VMWare Datacenters, one just for the disaster case. So we back up all servers on the main site, copy and replicate them to the other site.
Veeam 10 is installed on a fresh 2019 server with additional 2019 proxies, repositories are changed to REFS. Before that, we had a 2012 R2 server with v9.5 P3a.
Both the Backup server and all the proxies are VMs. We redesigned everything (jobs, configuration etc.) so we didn't use the config backup from the other installation.
The proxies are configured to use "disk hot add" method for best performance of course, without the option to fail back to network mode.

We figured out that when the proxies are replicating while being backed up the disk currently attached for replication is included in the proxy backup. Now we have
some big database servers which are replicated and need some time to be replicated (7x~1.3TB). The proxies then suddenly consumed up to a TB space instead of the 50GB C-Drive. After this happened there were many different and weird errors like "invalid snapshot configuration" on the replicated servers and bluescreens of the proxy servers from time to time. Just for clarification - we backup/replicate the database servers in a short interval, so there is no possibility to find a time window for the proxies.

Now we stopped backing up proxies and everything is just fine.

Is this a known behavior? I tought Veeam Jobs would know they are backup proxies and handle this different like the backup server which is not backed up until no parallel task is running. But I guess I am fatally wrong on this point. I don't really have a problem in not making a backup of our proxies, they are recreated very easily. But I don't remind having the same problem in v9.5. Maybe we just had luck.

[wishr]

Hi Thomas,

As of now the only supported option for backing up Veeam components is via configuration backup.

Thanks

[thomas_brun]

Alright, thanks for the info!

[spiritie]

We had a huge problem while backing up the proxies, at least on the disaster recovery site.

I have the same experience as this. Found out its best just not to backup the proxies because of the interference between hot-add process and snapshotting at the same time, it leads to all kinds of weird issues. Since all of our proxies it just plain windows servers with the proxy role, i am not too concerned not backing them up often. We've just made a job without schedule that we run once and now when no other jobs are running.

Had these issues in all versions of Veeam and all versions of ESXi.

[foggy]

Correct, if the proxy is used to backup VMs via hotadd, your best bet would be to either avoid backing it up or at least backup/replicate it separately. Note that such proxies cannot be backed up using CBT and cannot hotadd themselves, so will delay the entire job as will be entirely read over network.