Simplifying the Homelab

· seth's blog


In my post from late last year, I talked about things that I wanted to do in order to have a more resilient homelab. I think I'm getting closer to that goal.

I have put in a big order for some Zimablades in addition to some accessories for them to finally accomplish my goal.

What I hope this gets me #

Single Node Kubernetes #

I realize that this may seem to go against the "resiliency" I set out to achieve in my original projects post (what, no high availability?), but this is actually congruent with it. All of the apps that I use on my virtual cluster today is in configuration files. I can simply target a new Kubernetes cluster, and spin up the same apps again and everything would work out fine.

What that means though, is that as long as I have the configuration files I will always be able to recover my Kubernetes node to a working state. I will always be like, 20 minutes away from having a working git repo and Unifi controller. I have stripped out everything else that I don't need.

There are a few issues with this that I want to point out to the reader though #

For me, this change makes sense and will ultimately make it so I have a few key services without having to manage physical infrastructure that's heavier than a few pounds.

If you are wanting to start your own homelab, I came across a video in my research on the Zimablade that speaks to an idea I've had for a long time about homelabs. You don't need enterprise hardware to run a server. Anything you have, no matter what size, is sufficient. Follow the white rabbit.