proxmox on single disk in production

[pieterk]

So here's a thought. I have a few old Dell C6100 4 node boxes. Each node can address 3 disk drive bays. The aim is to run the nodes in a cluster with Ceph to have the virtual machine storage distributed. The aim is to only use KVM virtual machines and no containers. The disk division would be 1x OS disk and 2x CEPH OSDs per node.

Would running proxmox on a single disk in production be a legitimate option. Since they would all be clustered any failing system disk could potentially just be replaced and reinstalled. Does proxmox allow removing failed nodes easily from the cluster when a node is no longer available?

Does anyone around here run a cluster of proxmox nodes on single disks?

 

[hec]

I have installed system on single SSD.

The storage is coming from a NetApp MetroCluster with NFS.

Who cares if one node is failing as long as you have enough resources to restart the vms of the dead host.

 

[pieterk]

I have installed system on single SSD.

The storage is coming from a NetApp MetroCluster with NFS.

Who cares if one node is failing as long as you have enough resources to restart the vms of the dead host.

Have you experienced a node failure before? Is it easy to remove the dead node and maintain quorum with no access to the old node?