Ceph HA

[manues50]

Hello
I have two nodes in my cluster and was joined .
Now I was trying to enable ceph in both, I have successfully installed ceph in both servers, but the monitoring page is showing address: unknown, stopped for the second node

How can i troubleshoot the issues

 

[Stoiko Ivanov]

First - Ceph needs 3 nodes for a reliable operation in production!

else -check the journal of the host: `journalctl -r`
also read the ceph logs: '/var/log/ceph/*'

I hope this helps!

 

[manues50]

I have only two Poxmox servers and 1 TB each . and need to configure HA
Which one is best GlusterFS or CephFS ? or any other?

The intention is if one Proxmox server down the second will deliver the service (for both VM and Container )

 

[Stoiko Ivanov]

for HA to work properly you need at least 3 nodes (and an odd number is better) : https://pve.proxmox.com/pve-docs/chapter-ha-manager.html#_requirements

With 2 nodes I'd probably configure Storage replication with manual failover: https://pve.proxmox.com/pve-docs/chapter-pvesr.html

As an alternative you could add a qdevice to your setup instead of a third node - but this still would not take care of a reliable shared storage (but depending on your use-case storage-replication could be used)

I hope this helps!