Hi, can you paste the following output: ceph orch ls osd --export Maybe you have the "all-available-devices" service set to managed? You can disable that with [1]: ceph orch apply osd --all-available-devices --unmanaged=true Please also add your osd yaml configuration, you can test that with the --dry-run flag: ceph orch apply -i <osd_spec_file> --dry-run [1] https://docs.ceph.com/en/latest/cephadm/services/osd/ Zitat von Jan Marek <jmarek@jcu.cz>:
Hello all,
today I moved ceph to HEALTH_OK state :-)
1) I had to restart MGR node, then my old c-osdx hostnames goes definitely away and all of OSDs from old machines are now orchestrated by 'ceph orch' command.
2) I've updated ceph* packages on the osd2 node to version 17.2.6, then I tried 'cephadm adopt' command once more and voila! It works like a charm.
I will try to configure OSDs on the node 1 to adopt WAL and DB from prepared LVM... Maybe after upgrade to newer version of CEPH it will be OK?
Sincerely Jan Marek -- Ing. Jan Marek University of South Bohemia Academic Computer Centre Phone: +420389032080 http://www.gnu.org/philosophy/no-word-attachments.cs.html