Veeam Design Best Practices

[mmcd]

Hi,

We are a Telecoms company and have recently been tasked with using Veeam for backup and DR. We operate VMWare clusters in two seperate geographical locations, we have a 200Mb/sec link between the sites. One site will be active, while the other will be DR. Also, we have a Tape server unit, with around 8TB of disk and a tape library attached to this. As I am totally new to Veeam, I am looking for the following:

1. What is the best design for our scenario? Should we use the Tape server as the Veeam backup server and backup direct from SAN to this, then replicate a 2nd copy to DR? If this is correct or you disagree, please outline why.
2. If the above is correct, should we place a WAN Optimizer agent on a VM on each site? Or is the Veeam backup/management server capable of WAN optimization?
3. We are running SQL server mostly, as I am not a DBA, what considerations in the design and placement of management server should be taken for the transaction log backups?

Thanks
mmcd

[foggy]

Hi Martin, welcome to the forums.

1. Design mostly depends on your requirements, however what you're considering is pretty reasonable. You can use that server as Veeam B&R server and your primary repository and then tape out from it and also sync backups to a secondary location with the help of backup copy jobs. Also consider replication jobs to DR location for lower RTO in case of complete loss of the production site.
2. Veeam B&R has built-in WAN acceleration (which may not be required for 200Mb/s link though), so there's no need for any third-party tools.
3. Veeam B&R management server can be located anywhere, however there're some considerations regarding placement of log shipping servers, in case they are required.

[mmcd]

Hi Foggy,

Thanks for the quick reply. I forgot to mention that we do have a requirement for a 1 hour RTO(that is the aim anyway) and ideally an RPO of 30 minutes. So in relation to the replication part, is it best practice to replicate from the repository(disk backups) or from the production VMs to the secondary site? I assume from production as I doubt disk backups will be every 30 minutes.

I have used your free version of Veeam to migrate all our business to new hardware in our currently data centre. It made life extremely easy!

[skrause]

You will want to run your replicas from the production VMs, though you can do the initial replica seeding from a backup file to cut down on the impact on your production systems during the initial run as it copies all of the data across the WAN.

[mmcd]

Hi Steve,

Thanks for your input.

Finally, in the event of Primary DC failure, my Veeam B&R server will be gone, so how can I manage Veeam and initate my DR plans. Is it best practice to have a secondary B&R server on the DR site (such as a VM)? As this server is a physical unit, there would be no option for replicating it.

[foggy]

It's up to you to decide what works better in your particular case, however, with such a short RPO, remote replication indeed looks preferable. Having a secondary Veeam B&R instance responsible for replication at the remote location is indeed one of the best practices.

[mmcd]

Thanks for all the help, looking forward to getting some testing going with Veeam.

[skrause]

Having a B&R server in the target location for replication is how we have ours set up and it works well.