So I have a stuck server that won’t rejoin the cluster. Can I do a lxc cluster remove and get it to run container separately outside of the cluster? Or Perhaps somehow run containers manually via running it directly via lxc monitor.
I think the safe answer here is no. A clean removal requires there be no container or storage volume on the server being removed, so your only option would be forceful removal, which would leave that removed server unaware that its connection to the cluster was severed. It may or may not have access to the database at that point, so things would be pretty messy.
It really depends on exactly what stuck is in this case.
One last resort option would be to indeed kick it out of the cluster forcefully, then wipe its local database and certificates, re-init as an empty standalone server and then
lxd import the local containers into that empty LXD, but that uses the data disaster recovery method (
lxd import) so we’d normally prefer fixing the current situation over that.