Hi,
I have this same problem, and I thinks https://discuss.linuxcontainers.org/t/snap-auto-refresh-kills-cluster/2474 is related to this as well.
I have 2 separate clusters of 4 nodes each, in 2 different locations, and they were both unusable this morning.
On one cluster I solved by killing and restarting lxd (see https://discuss.linuxcontainers.org/t/lots-of-log-process-and-lxd-hangs/1592/2?u=sycolth)
(note that the lxc is not affected)
The commands I executed are:
systemctl kill snap.lxd.daemon
pkill lxd # or any other way to kill all those dangling lxd processes if any is left
systemctl start snap.lxd.daemon
systemctl status snap.lxd.daemon
On the other cluster I’m happy to test a more humane solution, if there’s any.