Consider the following limitations when backing up CSVs:
The nodes sharing the CSVs must be able to access the shared volumes. The shared volumes are located in the following directory:
<system drive>\ClusterStorage
Backups of nodes sharing CSVs cannot run concurrently. This approach ensures that the node that you are backing up has full control of input and output operations to the shared storage while the backup is in progress. For example, node A and node B are sharing CSV 1. You submit jobs to back up node A and node B. The backup of node B must start after the backup of node A completes.
When submitting backups consisting of VMs that reside on CSVs in Hyper-V systems, you must specify the Windows domain account for the Hyper-V systems in the Backup Manager Source tree. In addition, the Windows domain account must have backup operator and cluster administrator privileges on the Hyper-V system. This approach helps to ensure that backups of VMs that reside on CSVs in Hyper-V systems complete successfully. If you do not provide valid domain credentials for the Hyper-V systems, backup jobs will fail and generate the following message:
AE0603 Failed to create VSS shadow copy for the VM on the Hyper-V host machine.