help removing disk duplicate

[gfngfn256]

by manually deleting the corresponding folder in the storage images directory.

I would not be doing it that way.

The correct way from CLI to delete an unreferenced disk/s would be:

qm set <vmid> --delete unused<n>

# You could also with one command both unlink & delete the disk with:
qm disk unlink <vmid> --force <boolean> --idlist <string>

See qm manpage.

Pvesm uses "copy" and "move" as subcommands

Not sure what you mean by this, but AFAIK, pvesm which is the PVE Storage Manager would use the following to completely delete a volume:

pvesm free <volume> [OPTIONS]

#example from below docs to remove an image
pvesm free local:100/vm-100-disk10.raw

as shown in the pvesm manpage.

 

[Txalamar]

These commands work under the assumption that the disk is present in the virtual machine. In the state the VM was in, it would not have found any "unused" disk to delete.

It is clear that a rescan should have been performed first, followed by the usual procedure.

Just a note: Since I performed that single rescan, I have been unable to reproduce the error. Before that it happend every time a copy was left in the source datastore.

 

[gfngfn256]

It is possible that this was happening due to the storage being a NAS, but I'm not sure.

Just as a side point, I never tick that Delete Source checkbox, as I want to be sure everything goes as planned before any deletion, and I then remove the unused disk after that in the GUI.