In this latest recurrence, I can see that our new code path is ... affecting something, as ceph-mon/1 has no /etc/ceph/ceph.client.admin.keyring file. The hook will now try to go through the steps to generate that file 3 times before failing, so I once again suspect that there is *some* network issue between the containers containing the monitors. Going to continue to dig into the logs for the various units.
In this latest recurrence, I can see that our new code path is ... affecting something, as ceph-mon/1 has no /etc/ceph/ ceph.client. admin.keyring file. The hook will now try to go through the steps to generate that file 3 times before failing, so I once again suspect that there is *some* network issue between the containers containing the monitors. Going to continue to dig into the logs for the various units.