Method for removing hosts that are gone

currently trying to remove a host that is gone with "microceph cluster remove xxxx" will fail with

Error: Delete "http://control.socket/core/1.0/cluster/xxx": context deadline exceeded

I supose it is trying to contact the old host

A hacky way I have found to solve things is

microceph cluster sql "SELECT * FROM core_cluster_members"

followed by removing from database the approrpiate ID

microceph cluster sql "DELETE from core_cluster_members where id=3"

would be good to enhance microceph cli to directly force the removal.

Start with the `microceph cluster remove` command and reproduce the timeout when removing a host that is no longer reachable. Compare its behavior with the documented `core_cluster_members` SQL workaround, then verify that a supported force-removal path deletes the appropriate cluster member without contacting the missing host.

Written by the indexing model from the issue text.