Veeam Deployment Advice Needed

[wa15]

Hello all,

Need some deployment scenario advice here:

We have one main site and 5 remote sites located throughout the US with about 1TB of data to backup at each remote site. Each remote site has its own local repository, proxy, and wan accelerator. The approach we are currently taking is that we have all the jobs registered at the backup server at the main site and all are doing backups to their respective local repositories. Then we have backup copy jobs registered that copy those local backups to the main site.

We are looking at switching to a "distributed" Veeam topology by installing Veeam B&R on the proxy server at each remote site and have the local backup job AND the backup copy job registered at each remote site, instead of having all these jobs scheduled at our main site. Main reason for this is that currently doing FLR for a backup done at the remote site but registered at the backup server at the main site is NOT possible since the backup needs to be mounted to the backup server it's registered at which in this case would be the main site's backup server and mounting of a 100GB+ backup over a 20Mb connection times out.

We can manage all these backup servers via the Backup Enterprise Manager, though it is nice to manage everything from one application. My question is this, do you see switching to a distribution topology reasonable for the above backup needs? We have no replication going, so all that's in the picture is backups and backup copies. Will appreciate some feedback!

[frankive]

I had a customer with the same problem that you have; that you want to be able to do FLV on the remote site from the local repositories. I "solved" this by still letting the central Veeam Console on the main site do the job, but I also installed a "cold" console on each siste, but without secheduling job. I then just added the local repository. I can then restore from this repository locally. The only drawback is that each time I want to restore something i must go to infrastructure and rescan the repository so the console can register all the new restore points.

[foggy]

I also find Frank's suggestion totally reasonable. Having a dedicated Veeam B&R instance in each site just for faster restores is always advised in such cases.

[wa15]

Sorry for the super slow response!

@Frank, thank you for the suggestion. I don't quite see how installing a local cold copy of Veeam B&R will fix this though, since when doing an FLR, the backup job is mounted onto the Veeam server it's registered on. Unless we go in and import the local backups onto the local B&R machine. Can you please clarify? Thanks!

[foggy]

Yes, importing backup files into this local Veeam B&R instance is intended in this scenario (you should register your backup storage as the repository in that instance, backups will be imported automatically during rescan procedure performed on the repository).

[wa15]

Got it, thank you.

I know in another older thread Gostev mentioned that the possibility to mount backups to proxies is upcoming in a future version, so looking forward to that!

[foggy]

There are also a few other more creative workarounds, probably one of them suits you better.