I’ve formatted the drives on my single node Kubernetes instance.

I’ve removed the backups.

By removing the backups, I’m taking on faith that I can recreate anything so far from published blog posts as well as the unpublished topics in a private repository that I never quite wrote up.

I am rebooting my entire Kubernetes at Home series with the learnings so far.

Retrospective

This first run of the Kubernetes at Home series is hardly expansive with only nine blog posts before this one. The topics covered barely scratch the surface of what is there. Most of what I covered is setting up default installs of applications and sometimes tying them together for convenience. I didn’t show anything of how I was building applications on this cluster.

What went well?

• I learned the existence of more areas that I don’t know.
• I learned more of the command line tooling.
• I succeeded in building a single-node Kubernetes cluster, in spite of its flaws.
• I had fun.

What didn’t go so well?

• I didn’t blog about my custom software development. I only blogged about applications that were ready to install.
• My Kubernetes single-node hardware at home is almost too powerful. Developing on a powerful server at home can be a dangerous precedent to set when it comes to deploying to a cloud provider and keeping costs manageable.
• I depended highly on sane defaults from installed applications, including for the Kubernetes install itself.
• I neglected updating and upgrading installed applications.
• I deployed everything from my local desktop.

What actions may be taken to improve?

• Keeping a more frequent blogging-cadence.
• Being more minimalistic in what I install.
• Looking at K3s as a lightweight starting point. It looks like a clean starting point, and so less things I have think about.
• Trying out something like FluxCD in order to really do GitOps and to streamline local deployments and cloud deployments.

Kubernetes at Home: The Reboot

I’ve already started on season two of this series.

One key difference I can confirm right off is that I’ve installed openSUSE Tumbleweed server instead of Ubuntu Server. I’m still running Ubuntu on other machines, but what my working machines run is a separate blog post.