TL;DR: The last piece on this blog moved from external resources to a totally track-less state. Not for the server, but I can’t really kill that part. This post is a “quick” walkthrough of my journey.
Back years, long-long years ago, I had a server in a datacenter. Later she got a sister. The next few years went well, but then I was tricked by someone who could easily exploit my little experience in the business world and I landed on the streets for a while. I survived, no problem. I learned a lot, BUT I lost all my servers.
We were at the dawn of a new era, The Cloud. So as I managed to get on my feet again, I started to pay a little for virtual machines with limited permissions. I did not like it, so I moved to bigger companies and ended up at AWS… for like 2 months when I realized it will be freaking expensive. I tried a few providers and after long years I decided to stick with DigitalOcean because they always helped me if I asked them, they have reasonable prices and AWS independent. In the last few, years they improved a lot and they have a kind of clear vision of what they want to achieve and I like it.
As a side-effect, I started to rely on SaaS solutions like Disq.us, AWS S3, Google Login, GitHub, CircleCI, Travis, Google Drive, etc. Later I was not happy with that, but I could live with that burden. Easy without pain… Until all the big companies started to track everything heavily. I knew they did it before, but it was like “It’s fine”. But they started to track literally everything and everywhere.
I started to build up my own infrastructure (on DigitalOcean) and slowly I took over my data. Started to host my own Gitea instance with self-hosted Drone integration. It was a huge milestone as I felt I really own my own git repositories and their build pipelines. With small steps, I started to eliminate other smaller parts with Linx, PrivateBin, and a private PyPi repository.
At the same time, I started to feel the pain that I have two computers in my life (with work laptop). On top of that, I have a Raspberry PI too and a few virtual servers and my VimWiki setup did not perform well. It was the right time to fire up my own DokuWiki. The only pain-point was, I really missed the vim part. No problem, I’m an engineer, so I made a CLI tool so I can view, edit, and search my DokuWiki content. I can edit with vim.
Meanwhile. I went down deep in the federated universe with Mastodon, thank you @email@example.com. I started to think to start a new Mastodon instance so I can control everything there, like character limits or policies for example. In the end, I decided to use Misskey instead because I really liked the UI, the feature list it has, and how responsive and friendly @firstname.lastname@example.org is.
I started a new blog, this one, and decided to remove all ads from it, I’m happy if people are reading my blog, but I 100% sure, I don’t want them to be tracked. Most of them are most likely already has uBlock or something for it. Freedom over everything. But what about the comments? No one ever commented on any of my posts really or if it happened it was like 2 years later. I need a solution for that. So I found one, Commento did the trick. I tried to avoid using any kind of static content like images in my posts so I don’t have to use AWS or GHC, but I ended up with a few images. I felt remorse, but it was only 4 images, so yeah, I could simply build into my Hugo site, but then it would be painful to get that out of the git history.
But today, finally… I have my very own Minio server and all my images (yes all 4 images) are hosted there. Sounds a small step, but for me, it’s a huge one. Now I can add images to my posts without feeling guilty. So officially, if you visit this blog, only two domains serve the content
Both of them are owned by me. Please, if you find any other resources loaded from other sites, let me know, my contact information can be found on the front page.
TODO: the only thing left is a plugin for Hugo, so when I embed a YouTube video, I would like to fetch the preview image build-time and on-click load the video itself, so if a post has an embedded YouTube video in it, it does not call to Google and let them track you.