Ceph » Orchestrator

but doesn't that address only part of the problem. For example, any daemons that Ceph (not cephadm) knew about are still 'there' and down (mon, osd, mgr). Is this right? If so, are we expecting users to follow docs at that point? Also, when I tested this I saw stray daemons and stray hosts alerts from orch - and the host is no longer up and has been rm'd. Things like MONs or so will stay in the mon map and need to be removed manually.

Calling

ceph orch daemon rm <daemon>

for all daemons on that hosts doesn't work, as this requires SSH access to the host.

Calling

ceph orch host rm

only solves this partially as users still need to remove daemons manually, like ceph mon rm