I created a temporary ZFS storage pool + volumes in my LXD cluster, then created an ephemeral test container on each host.
I then did some disk speed testing (LXD pools+volumes & ZFS mounts were equally good - best RW mixed throughput), shutdown the containers (
systemctl poweroff) and tidied up.
In the process of tidying up I destroyed the ZFS pools behind the LXD storage pool and removed the LVs they used. Mistake.
lxc storage list showed the temp pool as in-use, so I couldn’t delete it. I restarted the hosts.
Now lxd is broken:
$ lxc cluster list Error: Get "http://unix.socket/1.0": EOF
# journalctl -b0 -oshort-precise | grep -i lxd ... normal startup Jun 08 06:43:50.903517 albans lxd.daemon: t=2021-06-08T06:43:50+0000 lvl=eror msg="Failed to start the daemon: Failed initializing storage pool \"temp-lxd\": Failed to run: zpool import temp-lxd: cannot import 'temp-lxd': no such pool available" Jun 08 06:43:51.045566 albans lxd.daemon: Error: Failed initializing storage pool "temp-lxd": Failed to run: zpool import temp-lxd: cannot import 'temp-lxd': no such pool available Jun 08 06:43:51.776278 albans lxd.daemon: => LXD failed to start Jun 08 06:43:51.777271 albans systemd: snap.lxd.daemon.service: Main process exited, code=exited, status=1/FAILURE Jun 08 06:43:51.777451 albans systemd: snap.lxd.daemon.service: Failed with result 'exit-code'.
Before I break it further:
- is there a way to remove the storage pool from LXD simply?
- will it restart if I recreate the destroyed zpool on all hosts