Backup Windows Failover Cluster with Storage Snapshots

[montenheimer]

Hello world,

since the newer version of the veeam agent, we are able to get use of backup-from-storage-snapshots.
Does anybody already have experience for backing up a windows failover cluster with the veeam agent while using bfss?

Szenario
Hypervisor: VMware
Shared Storage: Datacore
SAN: iSCSI

According to the technical documentation, the veeam agent computer and the backup proxy must have access to the LUNs over iSCSI san.
So we deployed a new windows server and use it as a backup agent proxy. The proxy has access to the LUNs. So far so good.
When we have "direct" connection to the san storage, we usually change the san policy to offline shared and disable automount, that windows cannot mount the luns and formats them.
Can we easily change that settings on a windows failover cluster, that we can have access to the LUNs on the cluster nodes (agent computers)? As soon as a failover occurs, does the node, to which the cluster failovered, still have access to its partitions after we disable automount?

Thank you in advance.
Markus

[RomanK]

Hello Markus,

In general, this feature does not care about cluster configuration at all. Firstly, the LUN clone is mapped to the active node (original disk location). Then the auto-recovery process is performed and the snapshot is marked as a transport snapshot. After that, it transfers to the proxy and is brought online forcibly for the block read.

The only note is that a snapshot must be taken in 10 seconds because of a cluster. QA team told me that for Datacore it should not be a problem for 5-7 LUNs but 20 LUNs could be too much.

Thanks

[montenheimer]

Hello Roman,

thank you for your reply. It seems obvious to me that the feature does not care about the cluster configuration.What do you mean with original disk location?

In the helpcenter its called:
2. Veeam Backup & Replication starts a backup job session and sends a request to the storage system to create a native snapshot as a new LUN.
3. After a snapshot LUN is created, this LUN connects to Veeam Agent computer to finish VSS operations and record storage metadata.
4. Veeam Backup & Replication moves the snapshot LUN to the backup proxy.

To my understanding, the new lun is mapped as a new disk. In this case, i would assume that this is the same behavior as doing backup from san (direct san access) where the proxy has direct access. When we are doing this, we have to disable automount and set the san policy to offline shared (on the windows proxy). Does this step apply to the agent configuration as well?

[RomanK]

Hello Markus,

The snapshot LUN is mounted on the machine with the original disk, i.e. where Veeam Agent was installed and VSS started.
On that machine LUN (disk) bring online. All volumes on it are set writable for auto recovery. After the auto recovery, all snapshot LUN volumes are set read-only. Then this LUN mounted on a proxy. This LUN will be online, but volumes would keep the read-only flags.

Veeam components on the proxy server are up to bring LUN online automatically despite the settings and automount.

Thanks