It might be a different issue, can you paste the entire stack trace from the mgr when it's failing? Also, you could go directly from Pacific to Reef, there's no need to upgrade to Quincy (which is EOL). And I would also recommend to upgrade to a latest major version, e.g. 17.2.8, not .0, otherwise you might hit bugs which have already been fixed in later releases. Zitat von Jeremy Hansen <jeremy@skidrow.la>:

I ran in to the “Error ENOENT: Module not found” issue with the orchestrator. I see the note in the cephadm upgrade docs but I don’t quite know what action to take to fix this:

ceph versions { "mon": { "ceph version 16.2.15 (618f440892089921c3e944a991122ddc44e60516) pacific (stable)": 3 }, "mgr": { "ceph version 16.2.15 (618f440892089921c3e944a991122ddc44e60516) pacific (stable)": 1, "ceph version 17.2.0 (43e2e60a7559d3f46c9d53f1ca875fd499a1e35e) quincy (stable)": 2 }, "osd": { "ceph version 16.2.15 (618f440892089921c3e944a991122ddc44e60516) pacific (stable)": 15 }, "mds": { "ceph version 16.2.15 (618f440892089921c3e944a991122ddc44e60516) pacific (stable)": 3 }, "overall": { "ceph version 16.2.15 (618f440892089921c3e944a991122ddc44e60516) pacific (stable)": 22, "ceph version 17.2.0 (43e2e60a7559d3f46c9d53f1ca875fd499a1e35e) quincy (stable)": 2 } }

ceph orch upgrade status Error ENOENT: Module not found

ceph config-key get mgr/cephadm/osd_remove_queue []

Thanks. I’m trying to upgrade this cluster all the way to current.

-jeremy