Homelab #25 Create a Proxmox Cluster with Tailscale

If you’re on the same LAN, just join via the GUI for a quick and easy setup. But my servers are in two different locations, so being on the same LAN is impossible. Still, I want to deploy a Cluster, and this post targets distributed nodes not on the same LAN

This is one of the hardest things I’ve done in my homelab journey. Mr. Thuan Bui also doesn’t have a post about this, so this one will be very detailed.

After removing the VGA, I installed Proxmox for Node PVE2 like in post 1, of course dealing with nomodeset

Then I pinned the PCIe port indexes using the command from post 24

Install Tailscale for Proxmox

Here I install directly on both PVE nodes that will join the Cluster

curl -fsSL https://tailscale.com/install.sh | sh
systemctl enable --now tailscaled
tailscale up --ssh

After running, you’ll get a link to log in to Tailscale

Configure Tailscale

I changed the IPs of the two PVE nodes to make them easier to remember

pve: 100.122.1.1

pve2: 100.122.2.2

Create Cluster

Go here to see if there’s already a Cluster. If not, click Create Cluster and fill everything in. For LAN, just enter the only LAN you have.

A legacy from post 1, this is the config from the very early days when I installed Proxmox.

To fix the communication port config (192.168.111.2), edit corosync.conf using the command below

nano /etc/pve/corosync.conf

In the highlighted part, I entered the Tailscale IP of the PVE node.

After filling it in, restart the corosync service with

systemctl restart corosync
pvecm status

Then the Cluster IP Address will be changed as desired.

Join the Cluster

Before joining, ensure a few things

• Both PVE nodes can ping each other and vice versa
• Their SSL must be the one signed by PVE itself

In post 14 I used Cloudflare Origin SSL, so I had to change SSL back to the one signed by PVE.

Start in Shell of Node PVE2

pvecm add 100.122.1.1 --link0 100.122.2.2

It shows like this. Check the Cluster and if the node is there, it’s a success.

Result

If you see this, it’s done!

Summary of some errors encountered

Delete Cluster on PVE

To delete a Cluster and redo from scratch, you can only do it via CLI; GUI has no delete button

systemctl stop pve-cluster corosync
pmxcfs -l
rm /etc/corosync/*
rm /etc/pve/corosync.conf
killall pmxcfs
systemctl start pve-cluster

Cannot connect to Proxmox WebUI

SSH into the Proxmox node fails, then use:

pvecm updatecerts --force
systemctl restart pveproxy

PVE2 joined the Cluster but is down

When PVE2 has joined but errors related to SSL cert or cannot connect to PVE2 while accessing the GUI, or PVE2 crashed/shutdown

pvecm expected 1
pvecm delnode pve2

Explanation: Proxmox nodes are equal, no master/slave. It’s democratic so you must force quorum to 1. To let PVE (node 1) have full control, post 26 will talk more about this.

References

Proxmox Cluster Manager Docs

Proxmox Cluster Manager Wiki

Tailmox

Proxmox Command Cluster Manager