Hi, I don't know how this happened but it seems second node's hosts file (/etc/hosts) was broken and "host-1" thinks itself as "host". Fixing /etc/hosts also fixed this issue. Thanks, Gencer. On 19.11.2020 17:33:52, "Gencer Genç" <gencer@gencgiyen.com> wrote:

Hi,

I ran those commands as usual:

$ ceph orch host ls Result is as expected by hosts names and addresses.

$ ceph orch ls Again, Expected result as before.

Then I started to upgrade via this command:

$ ceph orch upgrade start --ceph-version 15.2.6

It failed with attached logs. Please see logs.txt for detailed output. It says;

orchestrator._interface.OrchestratorError: DaemonDescription: Cannot calculate service_id: daemon_id='cephfs.host-1.shsruc' hostname='host'

This cluster have 2 nodes. Then I just ran same command again:

$ ceph orch ls Result is NOT as expected. I got;

...

...

Error EINVAL: DaemonDescription: Cannot calculate service_id: daemon_id='cephfs.host-1.shsruc' hostname='host' (See attached logs.)

If I restart "host-1" then i can execute "ceph orch ls" command. After I start upgrade it errors again. Interestingly I only have cephfs mounted in node "host". Not "host-1".

How can I fix this issue?

Thanks, Gencer.