We have a bunch of standalone ESXi 6.7.0 (Build 8169922) hosts with local storage. I've run into the problem that when I move a VM from host A to host B using an SCP transfer, and that VM currently has a snapshot, it will not start on host B. The error message is "Failed to power on virtual machine. Unable to enumerate all disks."
Now I've been running ESXi hosts for probably a decade and this always used to work without a problem, so I think this is a problem starting in ESXi > 6.5 maybe? I do nothing weird, shutdown the VM on host A, do an SCP of the entire VM directory to host B and boot it there. It only happens when the VM has a snapshot. If I SCP the same VM without a snapshot it will boot fine on host B. If I edit out the snapshot of the VM on host B after the transfer (by editing/removing references to the snap manually) it will also boot fine.
I have no clue why this won't work, because all the references for the snapshot should be in the VM directory (specifically the .vmsn and .vmsd file), not on the ESXi host. The hosts are the same ESXi version.
When I try to boot the VM and get the mentioned error, nothing is written to vmware.log in the VM directory. Also the ESXi host seems to write no logs about the error.
If I attempt to consolidate disks on the host B webinterface I get "Disk consolidation for VM has failed: The operation is not supported on the object". Also the snapshot manager in the web interface does show the current snapshots, but I can't delete them or create a new snapshot. All give a generic "failed" error.
Any help appreciated or confirmation that you have the same issue. If there are workarounds I'd love to hear them!