Your input for Veeam design

[omal422]

Hi, I'm looking for some input on how to best design our Veeam backup server set up on VMware. It's nothing complex, I don't think. I'm running through all the doco but thought someone could give me a well informed recommendation.

We have the following setup:
- Two sites (primary and secondary) with two vmware hosts per site
- One VCentre at secondary site
- Standard Windows servers - ie: DC, DFS-File/Print, TerminalServers, etc at Primary site
- Asynchronous Mirroring to secondary site for all servers (by SAN)
- Secondary site servers include DC, DFS, etc

We want to achieve the following to configure Veeam:
- Veeam server at Secondary site to backup servers from Primary site
- Backup repository duplicated at both primary and secondary sites and able to run restores from both locations (offsite storage)
- Restore VMs and file level at both sites - including if one site is completely down.
- Application aware backup jobs
- Backup Vcentre to both sites

I am thinking we don't need the Veeam Replica server backup (if I understood it correctly) because all our VM's are replicated to the alternate site using Asynchronous Mirroring. In the event of a server loss we can start up the alternate mirror.

A third level of offsite storage, such as cloud, could be added later if we need

How would you best design Veeam in this case ?

Thanks in advance for your input.

[ejenner]

One question... if you have a synchronized copy of your VMs at both sites... why do you want your Veeam B&R server sited at the secondary site using the primary site as source? Why not backup the synchronized VMs from the secondary site using the Veeam server at the secondary site?

[omal422]

Good question - but the asynch mirror is at a SAN level and the secondary volume is not accessible from VMware until we need it in a DR situation. The asynch mirrored VM's are not live.

[foggy]

I'd still backup locally in primary site (for fast operational restores) and then use backup copy to offline them. Remember the 3-2-1 rule.

[skrause]

I have found that running the VBR server that manages jobs on a VM in a non-production location, and having it use proxies and repositories in the production locations is very stable.

Since Veeam added the mount server and guest interaction proxy features, it is pretty easy to manage jobs "remotely" while keeping the components that are more network intensive "local" to the datacenter being backed up.

In a scenario where you have enough bandwidth to do storage replication as you mentioned, I don't see there being any reason why such an architecture wouldn't work well for your use case.