Windows File Cluster and CBT

[okrehan]

Hi folks,

we have a three node WSFC setup with ~45 data disks hosting nearly 100TB of data. We installed Veeam Agent v6 in latest version and also CBT driver to backup the disks.
Because we currently have some trouble with the cluster, all shares are active on one single node. The other two nodes have no roles assigned at the moment.
We backed up nearly 45TB of data in several runs as backup needs a lot of time. Because of a cluster reconfiguration we needed to switch the roles to one of the formerly inactive nodes but switched back a few minutes later without running a backup from the new node.
The next VB Agent run on the first node took multiple hours and rebuilded the CBT information as the existing information weren't valid anymore. This killed our backup windows because the job needed to READ 100TB of data just because CBT wasn't available anymore. (error: Changed block tracking map was changed for volume \\?\Volume{7bcdbaed-9849-489c-a70e-21a2dbbd9597}: current map f6f7d710-f338-11ed-9477-001999b75857, previous map c205ab3a-f0d2-11ed-9476-001999b75857 )

We added the clsuter in VBR as failover cluster so Veeam should be aware of possible role migrations and this shouldn't cause any problems but with our setup, it seems that each time we switch roles from one node to another, CBT is gone and we need to read all the data again.
I found the following note in VB Agent v6 manual (taken from https://helpcenter.veeam.com/docs/agent ... tml?ver=60):
If data blocks are changed on a volume while this volume is mounted on another Windows-based machine, during the next backup job session, Veeam Agent for Microsoft Windows will also read all data from the volume to create incremental backup.

Of course this will happen each time I switch a role to another node in a WSFC.
Does anyone else see this problem and found a solution? It's impossible to read multiple TB each time the cluster switches its roles.

Cheers,
Oliver

[HannesK]

Hello,
that sounds like unexpected behavior. Can you maybe tell us the support case number, or if there is none yet, please open a support case.

For situations where nodes are inaccessible or unplanned failover, the CBT map will be reset. But in your scenario, it should be no problem.

Best regards,
Hannes