Proxmox in One Evening

The hypervisor is not about running many machines — at the start you will run two or three. It is about making state cheap. Snapshot, try the scary thing, roll back in thirty seconds. Without that, every experiment carries the silent question “how long would a reinstall take?” and the honest answer stops you trying things. Which was the point of the lab.

The evening, in order

1. Flash the Proxmox VE ISO to a USB stick and install onto the machine’s SSD. Say yes to wiping it — this machine is a lab now. Give it a static IP when the installer asks; a hypervisor that changes address is a hypervisor you lose track of.
2. Open the web UI at https://<that-ip>:8006 from your normal computer. The self-signed-certificate warning is expected — accept it. Everything from here on happens in a browser tab.
3. Fix the repositories. A fresh install points at the paid enterprise repo and greets you with a subscription nag. Switch to the no-subscription repo in the UI (Repositories → disable enterprise, add no-subscription), then update. The nag is cosmetic; unpatched packages are not.
4. Make your first container. Templates ship built in — a Debian LXC container takes two minutes and a few hundred MB of RAM. Containers for Linux services, full VMs for anything that needs its own kernel.
5. Take a snapshot before you touch anything else. That is the habit this whole evening exists to install.

The static IP I set in the installer was inside the router’s DHCP pool, and three days later the router leased the same address to a phone. Both devices half-worked; the lab was unreachable in a way that looked like a Proxmox problem and was not.

The fix is one rule: hypervisor IPs live outside the DHCP range. Reserve the address in the router or shrink the pool — either way, do it the same evening, not after the collision teaches you why.