It has been a whole year since my first blog post. And to my surprise, I’ve kept to it. 😅 Sure, it’s just been a post a month for the most part, but I still managed to write a grant total of 31 posts. And it’s been fun!

The blog was born out of the fact that while tinkering with my Homelab, I’ve been reading so many niche blogs of other Homelab enthusiasts who did the same things I was trying to do. And I thought to myself: Hey, if I write about my tinkering, there might be somebody, someday, who finds one of my posts useful. And even if it’s just one person per post, it was already worth it.

And I’ve found that writing long form is actually fun! Ordering my thoughts, making screenshots or diagrams for illustration. It’s a nice way to spend a couple of evenings.

One downside of writing about my Homelab though: When trying to describe for other people’s benefit how I achieved something…I’m sometimes realizing that I have absolutely no idea why or how it works - just that it does. 😁

Look no further than my attempt at explaining OAuth/OIDC in the Keycloak article for an example. 😉

So did I achieve what I set out to do? Definitely! I’m still rather happy with the original trigger for creating a blog, my series about Netbooting Raspberry Pis.

Accessibility

I joined the Fediverse in August of last year. One of the cultural standards there is creating alt text for any and all pictures you post.

This made a lot of sense to me, and I also started doing it in my blog posts. But I soured on it. Whenever I sat down to start a post where I knew I would have a lot of pictures, screenshots and diagrams, I started thinking: Ugh, I will have to write alt text for this.

It got really annoying. I genuinely started to resent accessibility.

But it turns out: It wasn’t accessibility and the additional time it took to write alt text that was annoying me to no end.

It was the break in the writing flow. Having to stop and write (sometimes rather long) alt text every time I show a picture. Turns out that it’s not annoying at all if I skip writing alt text while writing the article and instead do it in one go during the final editing step.

So my tip: If you find yourself annoyed by writing alt text, keep it until the very end, instead of interrupting your writing flow.

What I’m still not sure about: Whether my alt text is actually any good. I often post screenshots of my Grafana graphs and dashboards. I’ve settled into a routine by now. I first describe the two axis, together with their minima and maxima. Then I describe the salient point of the curve(s). But I don’t know whether those descriptions are actually any good.

If you use alt text, please don’t hesitate to give me feedback!

Process

There isn’t really one. 😉 I’ve got a list of topics I want to write about, and then I just sit down and write. I’m pretty much a “write once” type of writer. So I write the entire post, put in pictures right away, and then never really change anything substantial.

After I’ve written everything I want to write on a specific topic, I read over the article one time, but that’s only to check grammar and see whether I have weird phrasing somewhere.

Then a final spell check with hunspell, writing the alt text for pictures, and that’s it.

The only change I’ve made to this process is getting into the habit of taking copious notes when I do something interesting in the Homelab, so I don’t run into too many problems with my rather spotty memory when I finally get around to writing the blog post on it. 😉

One weird thing about my process: I’m most productive at writing during train journeys. When visiting my family, I have a total of about 6 hours on trains, with only a single changeover. It’s this great train journey feeling of “There’s absolutely nothing I should be doing right now, besides sitting there and doing whatever”. No work, no domestic chores, nothing. So I’ve taken to writing on my train rides.

Tech

I’ve only just now realized that I have never described the Tech I use to run this blog, so here is a little description.

The basis is Hugo, and I’m using the PaperMod theme. I especially like that I can just run a local Hugo server on any machine and look at the rendered article while I’m writing it. Compared to some CMS system, I also like that my blog articles are under version control.

For delivery, I’m using my Homelab. When I push the blog repo to my Gitea instance, it triggers a Drone CI pipeline which runs Hugo and pushes the resulting page to an S3 bucket on my Ceph cluster. Then I’ve got my Traefik set up to forward external requests for my blog to that bucket and deliver the pages from there.

What’s next?

More articles! I’ve got so many articles on so many things in my ToDo list. And a couple more already in the drafting status. 😅

On the tech side, I would like to get more data about my blog. I know how many requests reach the blog each day through the analysis of my access logs, but that’s it.

I’m not hot on storing anyone’s private data or anything like that. But I do really like pretty graphs. Lots of pretty graphs. 😁 And I’m genuinely curious how many actual people read my blog, and what articles they read. Currently, I’m eyeing plausible.io for that purpose.

For articles, I’ve got several series I need to start/continue. My most fun one is a “Homelab History” series. I want to write about homelabbing from a different perspective than the tech tutorials I’m mostly keeping to. Starting with my first HP ProLiant server back in University to today. Not focusing so much on the apps running there, or the hardware they are running on, but rather on why I made the changes I did. I’ve been having a lot of fun digging through my past while putting together the outline for the series. Sadly, it’s been pretty difficult for the first couple of years, because I only started versioning things in the last couple of years. Before that, I only have my task management system, going through completed tasks and trying to dredge up some memory on why I did what I did. But fun nonetheless. 😄

The second set of articles I’m working on is the next installment for my Current Homelab series. While I’ve been pretty good about writing about the changes to the lab after I started the blog, there’s a lot of the current setup that hasn’t had an article, and that I think might be interesting for other homelabbers to read. I’ve already got my hardware described, as well as my storage setup. Next is a little bit about OS choice, and then the main event: My cluster, running the HashiCorp stack with Vault for secrets, Consul for service discovery and finally Nomad for container orchestration.

Finally, there are also a couple of tool articles I would like to write, for example about Taskwarrior, which has been running my life since 2014, or my experience running a Linux desktop as my main driver for 15 years.

Lots of stuff to come. 😄