[SOLVED] Ceph OSDs lost keys during PVE 5 to 6 upgrade

[VWNate]

Hello, I've recently inherited a proxmox cluster that is a few versions behind. In attempting to bring this system up to date and into compliance I'm afraid I've broken something and lost access to the ceph OSDs on host one of my three node cluster while updating from PVE 5 to 6. Any advice would be greatly appreciated.

 

[ubu]

Have you read the proxmox ceph cluster upgrade doku?

https://pve.proxmox.com/wiki/Upgrade_from_5.x_to_6.0

 

[ubu]

It might be easier and quicker to backup all your vms and containers, setup a new Cluster and import your vms onto the new cluster.

Not that you have to do it that way

 

[VWNate]

Have you read the proxmox ceph cluster upgrade doku?

https://pve.proxmox.com/wiki/Upgrade_from_5.x_to_6.0

Yes, the update to 6.x is actually complete, where I'm getting hung up is the Ceph Luminous to Nautilus update found here:

https://pve.proxmox.com/wiki/Ceph_Luminous_to_Nautilus

What I'm actually getting hung up on is the "Restart the OSD daemon on all nodes" step. The simple scan worked on node 2 and 3, but not on node 1.

 

[VWNate]

I ended up zapping my OSDs and rebuilding them, which has resolved my issue.