Best Practices VM backup for IP-less Exchange 2013 DAG

[jos3ph]

Environment:
4 nodes DAG - 1 active, 3 passive without lagged copy.
2 nodes in Site 1; 2 nodes in Site 2.
2 dedicated CAS VMs for mail access. Separate from DAG nodes.
IP-less setup.
Default cluster timing. No tuning.
Circular logging enabled.

Backup requirements:
- VM backup.
- Able to restore mailbox items from Explorer.

Questions:
1) Will backup one of the VM hosting passive copy work for Explorer restore?
2) What are the considerations and changes required to backup all VMs? Is there an advantage to do this?


Thanks a million.

[Andreas Neufert]

Please find my suggestions for IP-less DAG here:
https://andyandthevms.com/exchange-dag- ... plication/

On the same article, you find recommendations regarding cluster timing. Not only for Veeam but for all kind of VM operations such as vmotion and storage vmotion ... or snapshots you need to change this value to avoid cluster failovers during these operations. When you have multiple cluster failovers within minutes, the Exchange Index service will go nuts and will consum a lot of CPU resources for a long time (degraded performance).

Circular logging is not supported when you want to perform backups. See as well my best practice articel (link above).

1) Yes. I would just test it as well. I saw some configurations that had to select the databases in the explorer manually because the pointers were somehow broken. Please check and don´t forget to enable guest processing in the Veeam Job (indexing not required).
2) I would always backup all Exchange servers, even if that would mean that I need more backup space. In case of a disaster you can just use Instant Restore of the whole farm (or recover all servers) and the environment will come up. If you restore only one server you need to take extra steps so that the database will come up (start the information store service manually).