Hi,
Yesterday we started an upgrade of a cluster that was installed in 2017 and has been upgraded since. We were going to upgrade Ceph from Pacific to Quincy, and commenced by redeploying the mons, because the mons were still using leveldb, which is no longer supported in Quincy: https://pve.proxmox.com/wiki/Ceph_Pacific_to_Quincy#Important_Release_Notes
So we did. We removed the mon on proxmox03 and redeployed it. However, it will not join the cluster. It will stay stuck in 'probing'. We do see traffic back and forth, but that's about it.
The main difference by the look of it, are the names of the mons. Where the mons were called mon.0 and mon.1 back in the day, we now have the hostname in the mon-id. Could it be that the ancient mons (they really do date back to 2017) do not support that mon-name? Has anyone else ever seen this issue?
Yesterday we started an upgrade of a cluster that was installed in 2017 and has been upgraded since. We were going to upgrade Ceph from Pacific to Quincy, and commenced by redeploying the mons, because the mons were still using leveldb, which is no longer supported in Quincy: https://pve.proxmox.com/wiki/Ceph_Pacific_to_Quincy#Important_Release_Notes
So we did. We removed the mon on proxmox03 and redeployed it. However, it will not join the cluster. It will stay stuck in 'probing'. We do see traffic back and forth, but that's about it.
The main difference by the look of it, are the names of the mons. Where the mons were called mon.0 and mon.1 back in the day, we now have the hostname in the mon-id. Could it be that the ancient mons (they really do date back to 2017) do not support that mon-name? Has anyone else ever seen this issue?