Move Proxmox to another disk on the same machine?

[holckj]

I have a Proxmox installation, humming on nicely on a HPE Microserver with 4 disk bays. Proxmox is installed on a single disk, formatted as ZFS, but this disk is too small, and I also want some redundancy. So now I want to move the entire installation to two, larger disks, with ZFS Raid1 on the same hardware.

Is there a simple way to do this?

The Microserver is part of a small cluster with another server and a Raspberry Pi acting as qorum.

I have tried to install Proxmox to the Raid1 pool on the new disks. That's easily done, and the server can boot from the new pool. But I can't find out how to make this new installation join the cluster. I guess there may be some problems related to IP and mac-addresses - that the new installation still runs on the old hardware.

Any help appreciated

 

[aaron]

If I understand it correctly, you want to reinstall a node and join it again to the cluster?

Well, first you should disable the qdevice: pvecm qdevice remove. Then remove the node from the cluster.
Then, reinstall the node and join it to the cluster. Install the qdevice package on the new node before you set up the qdevice again.

Should there be problems regarding SSH keys with the new node, you can run pvecm updatecerts.

 

[holckj]

If I understand it correctly, you want to reinstall a node and join it again to the cluster?

Well, first you should disable the qdevice: pvecm qdevice remove. Then remove the node from the cluster.
Then, reinstall the node and join it to the cluster. Install the qdevice package on the new node before you set up the qdevice again.

Should there be problems regarding SSH keys with the new node, you can run pvecm updatecerts.

Thanks a lot. I tried this, but now I have in some way messed things up, I think.
Now pvecm won't run on the primary node in the cluster. I get this

# pvecm status
Cluster information
----------------------------
Name:                                68B
Config Version:                 17
Transport:                         knet
Secure auth:                      on

Unable to start votequorum status tracking: CS_ERR_BAD_HANDLE

I have one VM running on this node, but I can't back it up, because there is no quorum. How can I transfer this VM to another host, so I can reinstall everything on the primary node?

 

[holckj]

Answering part of my own question:

I managed to make most things look ok by doing:

# systemctl stop pve-cluster
# systemctl stop corosync
# pmxcfs -l

Now the node doesn't complain about missing quorum, and I managed to do backup with PBS with no problems. I also managed to make backups from the PBS accessible from another Proxmox server on the same LAN. So my plan now is to

1. Shutdown VM on old server
2. Backup VM with PBS
3. Recover VM on new server
4. Reinstall Proxmox on old server
5. Create new cluster on new server
6. Make old server part of the cluster

And then, hopefully, I can have a cluster with two nodes, making it easy to move the VM between the two.

 

[aaron]

Sounds like a plan. If I understand it correctly, the old server will be reinstalled and then added to the cluster right away, correct?

Then it should work. Adding a node to a cluster with VMs already on it won't work.