When I was dabbling a bit with CRIU/stateful migration I managed to completely destroy a container.
I was moving an ubuntu 18.04 container from a file backed storage (inside the lxd snap directory) to a zfs storage on the same machine.
The end result was an unusable container in the zpool and no container in the loopback storage.
I do not mind the the resulting container was unusable, but i did not manage to get my container filesystem. That was a simple test container, so no harm done, but I am a bit surprised that the original got deleted although the target does not come up.
How does lxc detect that the move was “successful”? Is there a flag to “delay” deletion?
I just realised that lxc move is probably just the lxd client command and not about LXC. I would guess that Incus has the same issue(?). So probably I should ask at the ubuntu LXD forum?