My homelab

I want to start keeping better notes about my homelab. Not because everything needs perfect documentation, but because I too often look at something later and think: why did I set it up this way again?

At home I now run quite a bit: Proxmox, VMs, storage, backups, network segments and various self-hosted services. Things get added regularly, and plenty disappears again too. If I do not write any of it down, only the end result remains. The reason behind a choice is easy to lose.

The base right now is a Proxmox cluster with seven nodes.

Why seven nodes?

Seven nodes is a lot for a home setup. There is no point pretending otherwise. One powerful server would use less power, need fewer cables and probably take less maintenance.

Still, I like not having everything on one machine. DNS, reverse proxy, storage, test VMs and regular applications are not things I want all on the same host. Not because that is technically forbidden, but because one mistake can hurt too many things at once.

I also do not use my homelab only to run services. I mainly use it to try things. For that, it helps to have a node where things are allowed to break without taking the whole environment with it.

Some machines have a fairly fixed role. Others are more for testing. I try not to mix compute and storage too much. Network services do not live next to experiments I do not trust yet.

That also makes maintenance less tense. If I want to update a node, change hardware or debug something, I can better estimate what will be affected. Not perfectly, but better than when everything runs on one box.

There are obvious downsides. More hardware means more power usage, more management, more places where something can fail and more cables behind the cabinet. For a purely practical home server, I would not recommend this.

For the way I learn, it works. I want to practice with separate roles, dependencies and failure domains. In that context, the extra hardware is not just overhead; it is part of the lab.

What I want to write down later

The parts around the cluster are just as important as the nodes themselves. The 10GbE backbone, VLANs behind pfSense, storage and backups on a separate Proxmox Backup Server all deserve their own notes.

I want to write short posts about those over time. Mostly to capture why I set something up a certain way, what turns out to be useful and what I will probably change later.