The only obvious difference between 19.2.2 and 19.2.3 I found was this commit: https://github.com/ceph/ceph/commit/32fae5ca4942f80a9604c6aa123442449594284e
cephadm: check "ceph_device_lvm" field instead of "ceph_device" during zap
And it's the only different line between those two versions [0],[1]. I created a tracker for this [2]. [0] https://github.com/ceph/ceph/blob/v19.2.2/src/cephadm/cephadm.py#L4273 [1] https://github.com/ceph/ceph/blob/v19.2.3/src/cephadm/cephadm.py#L4273 [2] https://tracker.ceph.com/issues/72513 Zitat von Eugen Block <eblock@nde.ag>:
Hi *,
I have a VM which I use frequently to test cephadm bootstrap operations as well as upgrades, it's a single node with a few devices attached. After successfully testing the upgrade to 19.2.3, I wanted to test the bootstrap again, but removing the cluster with the --zap-osds flag doesn't actually remove the VGs/LVs. This used to work just fine until 19.2.2. This is the command I used:
ceph:~ # cephadm --image myregistry/ceph_v19.2.3 rm-cluster --fsid {FSID} --zap-osds --force
There's not much to see in the cephadm.log after these sort of lines:
2025-08-01 09:04:02,103 7f3bf5206b80 DEBUG systemctl: Removed "/etc/systemd/system/ceph-6b501d0a-6ea3-11f0-a251-fa163e2ad8c5.target.wants/ceph-6b501d0a-6ea3-11f0-a251-fa163e2ad8c5@osd.0.service".
But it looks like the inventory output in json format, it's quite lengthy, so I'll spare you the output here. I can add it to a tracker though, if there isn't one yet. Has this already been reported?
Thanks, Eugen