Cluster Questions

[ballybob]

What is biggest disadvantage of using NFS over CEPH for a 3 node cluster?

Can you use NFS for an HA cluster where a failover will have no downtime?

How is memory state and running processes handled during this?

Does Proxmox primarily use one node for VMs, and then move them to the secondary node if there is an issue or is it balancing vms across all three nodes and moving them if a problem happens?

How does a vm handle HA failover in terms of IP, if for example I was using traefik to serve SSL certificates and have dns ips for these services, how do I ensure they can still be accessed after a fail over? Will it be using the same IP on the new node? (same LAN).

 

[leesteken]

What is biggest disadvantage of using NFS over CEPH for a 3 node cluster?

I think you have no redundant storage.

Can you use NFS for an HA cluster where a failover will have no downtime?

Proxmox HA always has a little downtime, so I don't understand the question.

How is memory state and running processes handled during this?

It isn't because HA kicks in only when the node on which the VM is running is detected as missing from quorum, and then the VM is started on another node.

Does Proxmox primarily use one node for VMs, and then move them to the secondary node if there is an issue or is it balancing vms across all three nodes and moving them if a problem happens?

How does a vm handle HA failover in terms of IP, if for example I was using traefik to serve SSL certificates and have dns ips for these services, how do I ensure they can still be accessed after a fail over? Will it be using the same IP on the new node? (same LAN).

I don't think it does those things and does not work as you expect. If you want seamless zero-downtime failover, it needs to be implemented at the software/service level. Proxmox cannot magically fix that for random software when the hardware of a node fails.

EDIT: Please correct me if I'm wrong about HA and Proxmox clustering (as I have little experience).