Failed to read disk over NBD transport mode

If everything is the same (access node, host, target etc.), then its likely that there is a stale lock or handle on the disk on the proxy/access node.

You can try this:

1. Suspend any Virtual Server Backups using the Proxy
2. Stop the CommVault Services on the Proxy
3. Navigate to C:\windows\temp\vmware-system
 
- Make sure System and Administrators have FULL Access or Rights to "vmware-system"
- Clear all files folders within "vmware-system" you may also need to Stop the VMware USB service as it can hold locks on files and not allow them to be deleted.
- Also if you see other instances of "vmware-system" in C:\\windows\\temp delete those only leaving the 1
 

4. Restart services
 

If that does not help, looking at the VixDiskLib.log file on the VSA proxy may help figure out what the issue is.

Some other suggestions for this sort of issue

1. Take a snapshot of the affected VMs - then use the “delete all snapshots” function to try do a disk consolidation
2. Storage vMotion the VM, and then move it back. This often clears up stale locks and processes which might prevent access to the disk. This fixes a LOT of strange VMware behavior with vmdk disk access

Thanks, I am still working on this. I noticed like you said, the VMs have snapshots from what looks like our app team doing software updates. Will this mess up commvault backups? 

No, existing snapshots on VMs is generally not an issue - it might become an issue if they are left around for a long time and accumulate a lot of changes. If the vMotion fails then perhaps that could be the issue, and you would need to delete the snapshots and consolidate the VM. I think that could be unlikely though.

We are trying to change the backups to Intellisnap backups.