Skip to content

Hello World!

What is it?

It's WOAHBase, a "thing" born of repetitive frustration. Stands for both

  • Where Only Awesome Happens ~ Base

  • Works Only At Home ~ Base

Cool, but like, Why is it?

In short, I'm always interested in the open-source and self-hosting my own services, and been containerizing dev-envs and workflows into docker images for a few years now. About time it is FOSS-ed out so everybody can benefit from them.

Feel free to use/tinker/break/fix any of the images, get in touch if you're looking to contribute time, if you're building something awesome and the images (or code) have saved your time, consider giving a shoutout.

The Long Story

Oh boy! I do hope you got time to read this wall of text.

The Problem

We're all familiar with the woes of trying to set up our machine just as we like it. But installing / configuring / updating and cleaning up seems to entail too much work sometimes. Even with automation, you still have to monitor it, and react to its hiccups. And when you get on to my level (please don't, as of 2022, homebase is running 50+ containers on multiple and multi-arch devices), sometimes it does seem too big to sink a tooth in. So circa June'17 I went looking for reproducibility, semi-permeable mutability but most of all, redundancy and reliability.

Enter Docker.

Docker containerizes all important processes neatly into separate layered images. Is quite well portable as we're moving whole OS images. Running is a breeze as long as its configured properly at runtime. Still, image size becomes an issue sometimes.

Enter Alpine Linux.

Now we have smaller images, the mini-rootfs is about 5mb, its blazing fast and pretty secure, deploying and going live in seconds. So, what seems to be the issue now?

After a while, you end up with a bunch of those images. To the point it becomes tedious for one person to build/maintain them.

There was Travis.CI(.org)

They allow building Open-Source Projects with their spare buildpower. How cool is that? Now all our base images can build there, only the special sauce needed, we can add it on the fly or customize it before running.

Thank you Mr. Travis. :)

Thus, from the locally hosted repositories, The code went to live at Github. And,

WOAHBase was assimilated.

And it was all fine and dandy at my homebase. Things seemed to work good for a while, no issues too hard to handle. But then, the dark times started around Dec'19. Travis announced the end of builds for free on April 1, 2020. It's cool though, it was too good to last anyway, and pretty fun while it did.

Going Dark

By the time the builds stopped, I already had a somewhat convoluted-but-works-well buildbot setup going for me, mainly used for small tasks, daily backup or other periodic jobs. So I moved all the builds right there, running on a i5/32G Intel NUC. Keep in mind though, I'm just one guy, no close to Travis.CI, so as I got busy with other stuff in my life, the update frequency soon fell to zero, and after Docker Hub announced they were going to put a limit to their pulls, setting up my private registry seemed the best way to go. I already had my own private git repository that goes back to 2012. So eventually, this tiny part of Github/DockerHub called WOAHBase kind of became deserted.

Back to the Light

So that's how it was for a few years, all-the-while I was feeling kind of an itch inside that I'm only building these images for myself, but running maintenance for the whole site along with 80+ images seemed like a mountain-size chore. Since about Q1/Q2-2024, I success(?)fully guilt-tripped myself into deciding that a change was due. And now that I'm considering job-hunting again, I could finally allocate the time to freshen up the the whole project. In the end,

Everything that follows, is a result of what you see here. Thank you for your attention.

Why not use Service-X for free automated builds?

Took a lesson from Travis.CI, and stopped mooching off of others' spare CPU power. My buildbot works fine, and albeit slow, I'd rather have something dependable that I can configure once, then just pay for to upgrade and handle heavier loads in future, compared to dealing with the headache of moving ~100 builds onto different platforms every few months just because it was free (only in monetary value) before, and is not anymore.

I keep dreaming about having enough spare cash someday to run a semi-public-yet-isolated build cluster that can handle all these jobs, but until then, there is nothing stopping anybody from forking an image-builder and setting it up to build under their own responsibility. And if you do end up doing that, let me know, I might link it in the docs. ;)