That is likely the same issue as the other thread I linked to had. During snap refresh the mount namespace seems to lose track that the pool is already mounted and tries to mount it again, even though it is mounted because the other containers are running.
It is likely causing this check to incorrectly return false, and trigger a mount attempt.
https://github.com/lxc/lxd/blob/master/lxd/storage/drivers/driver_btrfs.go#L295-L298
@stgraber have you seen anything like this before?