Recently I reinstalled both of my personal machines and with that came the process of regenerating all the keys and credentials I need - I backed them up but I find it’s a good opportunity to do a refresh. One of those keys is to access my home Kubernetes cluster.

The cluster has been running on Talos and one of its advantages is that I can keep the config for the cluster in git and use it to recover the machine. A side effect of that is it becomes really easy to regenerate the kubectl config.

Well it finally happened. After just over a year of its current incarnation and about another year on top of running without major downtime, I finally managed to break my Kubernetes (K8s) cluster to the point where I need to rebuild from scratch.

Thankfully I still have access to my volume data so it should be recoverable, however I am using this opportunity to take stock of things and learn some lessons to make my life easier in the future.

I have recently been learning and using Kubernetes in my homelab as a means to manage a lot of self-hosted apps. Overall I’ve been pretty happy with it and it has helped simplify my setup.

The first question to answer is why to use a system such as this in the first place. In the last 10 years or so there has been quite a shift in the way that people deploy and run their applications and resources. In the beginning there was bare metal where you would have one physical machine with one OS that hosted multiple applications. Then virtualisation came along and allowed people to segregate those applications so that they had their own OS stack for each, isolating the resources which caused less conflicts and allowed for easier management. More recently we’ve got containerisation platforms which make far better use of the available resources by sharing low level components with the host OS. Tools such as Docker have given an easy means of packaging and deploying containers which also helped contribute to their popularity.

When you enter my flat, you can hear the faint hum from the cupboard. Peering into it reveals a stack of networking and server equipment. When people see this, they often ask: What’s this all for and why do you have it? Homelabbing, as it’s known, is a hobby where people have enterprise level networking infrastructure and a variety of servers in their own home. It’s something that I had been doing on-and-off for about 5 years but more actively for the last 6 months (now that I have my own home). In this post I want to discuss the reasons behind why I do it; who knows maybe I’ll get some others interested! In a series of other posts I’ll discuss anything of interest in my setup.

As part of my final year project at university, I created a basic implementation of a Web of Things network (more information can be found on my project page). It was my intention to carry on developing this project once I graduated, and evolve it into a much larger system than merely a proof of concept as it currently stands. I’m a bit delayed but now is the time to begin making those changes. This blog post will outline the key things I want to add over the coming months, and I’ll check back in with progress when I can. Note that these are all changes on the servient, the client will need to be updated in order to suport these however this will come later.