Tailscale (Headscale), and other server thingys

I have used Wireguard for a long time to connect to my webserver from abroad, and connect to my desktop computer similarly.

Wireguard is great, but it seems to have most of its strength in static configurations, with a few nodes connected to one hub.

So recently, when I wanted to have all my machines interconnected on a pseudo-local network, I looked to Wireguard, and found it lacking in the niceties department. Sure, I could take time to painstakingly create every wireguard configuration for each device, but what would I do if for example devices switch networks? It’s all just a hassle.

I had heard of Tailscale before, but I was turned off by the fact that some of their code is not open source (the protocol is, but the control server, and some of the clients, aren’t). This is a big deal to me for auditability purposes.

What I had not heard of is Headscale, a reverse-engineered alternative control server for Tailscale. (Also, need I mention open source?)

With this, I can potentially excuse the Windows client not being OSS, as long as I am hosting my own control server.

So, that’s what I did. It took me several hours to figure out that I fat-fingered an IP address in my configuration, but after I figured that out, I now have a functioning tailnet hosted by me, myself, and I!

I can connect all my Linux (and some Windows for work) machines under one pseudo-network. This unfortunately means I’m going to say goodbye to my Wireguard configurations, but that’s ok. I think Tailscale is an improved abstraction layer over Wireguard.

Now if only someone would make FOSS alternative Tailscale clients for Windows / iOS / etc. Let’s get a fully open stack in here :)

Anyways, that’s all. It has been a while since I’ve gushed about anything homelab related, so maybe this is a nice cleanse from the current state of software development.

I’ll have another blog post about my Fire Tablet mainlining up soon enough!