Branch-Office Backup/Proxy-Server recommendation ?

[rolandk]

We are using older Fujitsu RX100 Servers in several Branch-Offices as Veeam Proxy and Repository Server.

As they are old, consume too much power and also only have slots for 2 Disks with maximum 2TB only (which is ok for now, but as those offices grow will not scale), i´m curious which replacement is recommended by others.

As we have only two ESX Servers in every branch office, i`m not sure if it would be worth the effort to put two extra backup-proxy servers on each ESX to be able to use a NAS as a target. Despite that, i don`t have a cloue how to use instant restore with NAS and using NAS will also increase network traffic, as you cannot run the proxy agent on the NAS. Furthermore, i read that backing up to CIFS is unstable or at least not recommended.

So, what type of machine do you use for backups in your branch office ?

What about avoiding the extra server and put a "mirrored" virtual veeam backup/proxy infrastructure on top of the ESX servers, so putting veeam backups to local datastores on BOTH ESX servers (maybe with raw disk mappings to mitigate the risk of unintentionally deleting the backup-VM(s) ) ?

[Vitaliy S.]

Hello Roland,

As we have only two ESX Servers in every branch office, i`m not sure if it would be worth the effort to put two extra backup-proxy servers on each ESX to be able to use a NAS as a target. Despite that, i don`t have a cloue how to use instant restore with NAS and using NAS will also increase network traffic, as you cannot run the proxy agent on the NAS. Furthermore, i read that backing up to CIFS is unstable or at least not recommended.

What network connection link do you have to the NAS box? 10GbE? You're correct by saying that you will increase the network traffic, however given that you have a good network connection to the host and NAS box from the proxy, it should be fine to run a vPower NFS server on the proxy. This vPower NFS server will be used in Instant VM recovery process. Using CIFS as a target for the backup job, can indeed cause some issues related to CIFS itself, however many customers are successfully using this kind of setup now.

What about avoiding the extra server and put a "mirrored" virtual veeam backup/proxy infrastructure on top of the ESX servers, so putting veeam backups to local datastores on BOTH ESX servers (maybe with raw disk mappings to mitigate the risk of unintentionally deleting the backup-VM(s) ) ?

It is possible, but if you lose your ESXi host, you will not be able to reach your backup storage to run a restore job. This is the main drawback of this approach, but these days installing ESXi server takes very little time and this might not be the issue for you (depends on the RTO policy in your branch offices).

As they are old, consume too much power and also only have slots for 2 Disks with maximum 2TB only (which is ok for now, but as those offices grow will not scale), i´m curious which replacement is recommended by others.

Old doesn't mean bad or something that should be immediately replaced. I would consider replacing these servers once bottleneck stats start finger-pointing to the target storage. As to possible recommendations, then let's see what other community members can say on this.

Thank you!