I’ve always appreciated the importance of integrity in my life and of course that also applies to the life online. It wasn’t until around 2018 I actually started taking it seriously and took action to get away from Big Tech like Google, Meta (Facebook at the time), Microsoft etc. At that point I went from Windows to Linux and after a while I felt the need for a place to document the things I was doing and this website was born.

As probably noticed it didn’t go as planned as I haven’t written as much as I wanted to, it has been a post here and there. I’ve always had the ambition to write more than I’ve done though.

Jekyll

When I first started this site, in 2020, it was generated with Jekyll. I tried a lot of different solutions and ended up with Hugo. I was never really happy with Hugo, but stuck with it because the annoyances weren’t big enough to actually switch again.

Now it’s time to make the switch again. This time, back to the roots - Jekyll.

Self host

I started hosting my own services around 2019 I think. At the time I had everything on a VPS. In late 2022 I bought my first server to start my homelab at home on my own hardware. It was kind of a stupid buy to be honest. I did build my own server out of consumer hardware and it was not the cheapest option. On that server I installed Proxmox and installed TrueNAS in a VM and ran a couple of other VMs for other stuff.

After a while I wanted to run TrueNAS on bare metal so in the summer 2023 I bought my next server, a HPE ProLiant ML110 Gen10 which I installed TrueNAS on and that server is still up and running. Since then I’ve acquired some more hardware.

In 2026

Now, in 2026 my homelab looks like this:

• My router is an old computer I had laying around. I bought a X540-T2 10G NIC for some better networking
• A Proxmox node is running on a Minisforum UM760 Pro I bought in 2023. That mini PC is running all my VMs and containers
• The ML110 is still running as my NAS, which I’m on my way to switch out for a UGreen NASync DXP4800 Plus with 32 GB RAM and TrueNAS installed on it. I bought the UGreen NAS 6 months ago but it’s still not in use. I’ve been procastinate it for too long now
• A RPi running Pi-Hole
• A couple of switches

What I’m hosting

• UniFi controller
  I have a UniFi switch and a UniFi AP so I’m hosting a UniFi controller as a container in Proxmox.

• Linkwarden
  A bookmark/read it later manager. Linkwarden replaced Wallabag which I’ve a post about. Running as a VM in Proxmox.

• Home Assistant
  For my smart home stuff. It’s not that much yet. Running as a VM in Proxmox.

• Gotify
  Receive notifications for stuff in the homelab. Running in a Docker container in Proxmox.

• Miniflux
  A RSS feed aggregator. Running in a Docker container in Proxmox. I’ve a post about installing Miniflux. I’ll post a new one for installing with Docker one day. Running in a Docker container in Proxmox.

• May
  Track fill ups, costs and other stuff related to my car. Running in a Docker container in Proxmox.

• Vaultwarden
  Selfhosted version of Bitwarden password manager. Running in a Docker container in Proxmox.

• Maloja
  Self hosted music scrobbler, like LastFM. Running in a Docker container in Proxmox.

• IT Tools
  Some handy tools. Running in a Docker container in Proxmox.

• Wireguard
  VPN to be able to always use the benefits of my home network and access all my services wherever I am, like ad block and stuff like that. Running in OPNSense.

• Redlib
  Reddit front end. Running on a VPS.

• SearxNG
  A meta search engine. Running on a VPS.

• Snikket
  XMPP server. Running on a VPS.

• Nextcloud
  I guess most people know Nextcloud. I mostly use it to sync my mobile phone photos, contacts and calendar. Running on a storage box in Hetzner.

• Pi-Hole
  Ad-blocking and DNS for my network. Running on a Raspberry Pi.

• Syncthing
  Syncthing is used to sync files between devices. Currently running on my TrueNAS system.

• Traefik
  My reverse proxy.

I have some more stuff running but that’s not in use at the moment. I also probably forgot some things.

Obtainium

Obtainium comes to rescue. I found this project a while ago and I love it. With Obtainium you download the APKs straight from the source, meaning you don’t have to wait for F-Droid to build the app for you…in most cases. The thing is that the developer of the app you want to donwload have to offer an APK from somewhere to download, either from the git repository or from a website. In the cases where they don’t you can also supply the F-Droid APK, which is really neat, but in those cases you wont be able to avoid the F-Droid build process.

They also provide a site where you can add Crowdsourced app configurations for popular apps to Obtainium.

No more F-Droid

I’ve moved most of my installed apps to Obtainium now and using that almost exclusively. I still have F-Droid installed on my phone to be able to browse for apps, but I do download and install everything via Obtainium.

It’s a very neat app. You can read up more on it on their Github, website and their wiki.

How is it done?

First we need to enable GitInfo in our config file hugo.yml:

enableGitInfo: true

Next we have the actual code which will display this in the single.html. For me, this is in layouts/partials/metadata.html.

In the top I define the variables:

{{ $date := .Date.Format "2006-01-02" }}
{{ $lastmod := .Lastmod.Format "2006-01-02" }}

And to show the date:

{{ $date }}
{{ if ne $lastmod $date }}
  | Updated: {{ $lastmod }}
{{ end }}

Right now I need to add lastmod: with the same date of the post to not show the date if it’s the same day. If I comment it out it will show for some reason. I didn’t care that much about to start dig for a solution.

How it looks

This post is updated, just to see how it looks. It doesn’t look good, but I have no better solution to it right now.

Now all I have to do to deploy my changes is add them to git, commit and push, everything else is handeled by the git hook.

More on that in a bit.

Assumptions

Before starting there are a couple of things that needs to be checked.

You need to have a…:

• bit of knowledge of git, atleast know what it is and how it works
• Hugo site set up
• server with ssh access to host your site on
• web server installed and working on the server

When all of that is checked, we continue to set up our repositories.

Set up the repositories

On the local computer inside the folder of my website development I created a git repo:

git init

On my server, where I’ll publish the website I created a git bare repo:

mkdir git/odd.unover.se.git
cd git/odd.unover.se.git
git init --bare

The script

The script I created looks like this:

#!/bin/bash

DOMAIN="odd.unover.se"
GIT_REPO="$HOME/git/${DOMAIN}.git"
BUILD="$HOME/git/build"
PUBLIC="/var/www/$DOMAIN"

set -e

rm -rf "${PUBLIC:?}"/* "$BUILD"
git clone "$GIT_REPO" "$BUILD"
/usr/bin/hugo -s "$BUILD" -d "$PUBLIC"

The script is very simple, but I’ll break it down line by line.

We first create a few variables: the domain we’re using, the path to the git repo, the directory where we build the site and the location where the site will be served from:

DOMAIN="odd.unover.se"
GIT_REPO="$HOME/git/${DOMAIN}.git"
BUILD="$HOME/git/build"
PUBLIC="/var/www/$DOMAIN"

Before we start doing anything we are going to clean up both the build and public directories, just to avoid potential issues:

rm -rf "${PUBLIC:?}"/* "$BUILD"

When we’re pushing from our local development environment, we’re going to push to, in my example, odd.unover.se.git that we created earlier.

Next up is to clone the git repo (odd.unover.se.git) to the build directory:

git clone "$GIT_REPO" "$BUILD"

Now we can build the actual site with Hugo. We’re using the build dir as source and the public dir as destination:

/usr/bin/hugo -s "$BUILD" -d "$PUBLIC"

Deploy

Now on from the git repo on the local computer we need to add the repo:

oux@odd.unover.se:git/odd.unover.se.git

Where oux is the user. odd.unover.se obviosly the domain. :git/odd.unover.se.git is the path to the git repo on the remote server.

git add .
git commit -m "commit"
git push origin main

Every time I push the changes I’ve committed, it will deploy it on this website as well as give me the version control.

When I started this site almost 5 years ago I had an idea of me writing a lot and document a lot of things I do, but it never really happened. I still want to do it, but I spend a lot less time with the computer these days.

This site have had more designs than posts and I’ve tried out way too many static site generators and domains over the years. I’m never satisfied and redoing stuff all the time. I’ll do my best to post more stuff in 2025 instead of just redoing everything all the time.

New changes

One point in time I messed up my git repo. I can’t really remember what happened, but that means I have to create a new repo. The sad part is that all the history is gone as well as a fair amount of posts, but hey… It could be worse.

I’ve rebuilt the site from ground up and really like how it starting to look. Simple, minimalistic and easy to maintain but still very flexible and powerful. I’m still using Hugo and probably will keep doing so, but I’ve created a new site and the theme myself.

Another thing that we are going to fix here is to automatically start X when we’re logging in to our user.

Auto login

There is a simpel fix to this. We need to make a change in the file which controls the startup of TTY1. To be on the safe side, we can create a copy of the file.

cp /etc/sv/agetty-tty1/conf /etc/sv/agetty-tty1/conf.bak
vim /etc/sv/agetty-tty1/conf

In there we can see an if statement and in there we find a variable called GETTY_ARGS that we need to add an option called --autologin followed by our user name.

if [ -x /sbin/agetty -o -x /bin/agetty ]; then
	# util-linux specific settings
	if [ "${tty}" = "tty1" ]; then
		GETTY_ARGS="--noclear --autologin oux"
	fi
fi

Start x

To skip the need to type startx everytime the computer starts is a nice thing to have as well.

This time we need to add these lines below to our ~/.xinit file:

if [ -z "$DISPLAY" ] && [ "$(fgconsole)" -eq 1 ]; then
	exec startx
fi

What to do?

All my machines are always encrypted and I’m the only user on them, so for me it’s just redundant to have to login to the user each time since the encryption needs a password to decrypt. Instead of having to login twice I just enable autologin for my user.

Void Linux uses the runit init system and to enable autologin we need to change the tty service file.

Open /etc/sv/agetty-tty1/conf and add the -a flag and your username after any other arguments in the GETTY_ARGS variable

doas vim /etc/sv/agetty-tty1/conf
GETTY_ARGS="--noclear -a <username>"

Now we will always be logged in automatically when we start up the machine.

x

We still got one problem though, x is not starting automatically, so even if we’re logged in, we’re still in the TTY, and I would prefer if my system could startx for me, so we have to put the following in to .bashrc

if [ -z "$DISPLAY" ] && [ "$(fgconsole)" -eq 1 ]; then
	exec startx
fi

When this is done, we’ll be logged in automatically and the system will start x for us.

Wallabag is a free and opensource Pocket-like software. It’s main function is a ‘read later’-service which grabs text and pictures from links you add and Wallabag shows it clutter free and without all the bloat on todays web, like the Firefox Reader View. I mainly use it as a bookmark service though.

One thing that i now realize i’ve missed almost most is Wallabags ability to share articles you save with a public link so that you can share articles without all the mess of most webpages.

It’s of course possible to selfhost, but there’s also an Official hosting service for a small price tag if you’re not up to selfhosting it. Another option is to find someone who hosting a public Wallabag instance for free.

One of the biggest reasons i switched back from Shaarli to Wallabag is some annoying things when sharing from the Android app Shaarlier, which made me almost stop using Shaarli after a while. Another thing i really like is that I’m able to save articles from my Miniflux to Wallabag with one click only.

After some minor database issues during the installation I’ve finally installed Wallabag again and I’m very happy with it. It looks way better now than it did the last time i used it.

Taking a screenshot

The command for take a simple screenshot is

import -window root screenshot.png

This will take a screenshot of the whole root window, which is your screen.

How I use it

I have a script that I have a keybinding to that opens a rofi menu with different options.

#!/bin/bash

# options to be displayed
option0="screen"
option1="area"
option2="window"

# options to be displyed
options="$option0\n$option1\n$option2\n$option3\n$option4"

selected="$(echo -e "$options" | rofi -lines 1 -dmenu -p "scrot")"
case $selected in
    "$option0")
		import -window root /hdd/pix/prtsc/"$(date +%Y-%m-%d-%T.png)" && notify-send 'Screenshot' 'Done';;
    "$option1")
		import /hdd/pix/prtsc/"$(date +%Y-%m-%d-%T.png)" && notify-send 'Screenshot' 'Done';;
    "$option2")
		import -window id /hdd/pix/prtsc/"$(date +%Y-%m-%d-%T.png)" && notify-send 'Screenshot' 'Done';;
esac

1. The first one just takes a screenshot of my whole screen in .png format and set the date and time as filename. After that it sends me a notification with dunst that a screenshot is captured

    import -window root /hdd/pix/prtsc/$(date +%Y-%m-%d-%T.png) && notify-send 'Screenshot' 'Done';;
   

2. In the second option it is almost the same, but without the -window root part. This makes it possible for us to choose which part of the screen we want to capture using the mouse

    import /hdd/pix/prtsc/$(date +%Y-%m-%d-%T.png) && notify-send 'Screenshot' 'Done';;
   

3. And the last one is with -window id. With this option we can chose a specific window to capture, this also using the mouse

    import -window id /hdd/pix/prtsc/$(date +%Y-%m-%d-%T.png) && notify-send 'Screenshot' 'Done';;
   

I do recommend to read up on the options that’s available. There is a lot of options you can use if you look at ImageMagicks website or the man page.