sudo dpkg-divert --divert /etc/apt/apt.conf.d/20apt-esm-hook.conf.bak --rename --local /etc/apt/apt.conf.d/20apt-esm-hook.conf

This will add .bak to the end of the file and should survive updates.

I installed virtualbox and set up a vm with shared host adapter for being able to easily access it via ip. On that ubuntu vm which I access from outside via SSH I installed docker as documented on the official docker documentation and gave my user proper rights to use docker. The vm has proper internet access and can run docker containers with ports mapped to the shared host adapter.

From my mac I installed docker via brew (not desktop) and added a context for the vm:

docker context create vm --description "local ubuntu vm" --docker "host=ssh://manuel@ubuntu"

The “ubuntu” hostname was added to my /etc/hosts with the configured ip of the vm and manuel is my ssh user on the vm.

With the new context created the context list looks like this:

➜  ~ docker context ls
NAME                TYPE                DESCRIPTION                               DOCKER ENDPOINT                                KUBERNETES ENDPOINT   ORCHESTRATOR
default *           moby                Current DOCKER_HOST based configuration   unix:///var/run/docker.sock                                          swarm
desktop-linux       moby                                                          unix:///Users/manuel/.docker/run/docker.sock                         
vm                  moby                local ubuntu vm                           ssh://manuel@ubuntu

After switching the context to `vm` I can easily work with docker running on the vm.

docker context use vm

With the use of context I was able to come around docker-desktop completely. So the argument that docker isn’t usable anymore on a non linux machine is busted.

snap list | awk '{print $1}' | xargs -rn1 sudo snap remove

After letting this run a few times i tried to uninstall the remaining snaps manually with

sudo snap remove <name>

Some refused to uninstall but I continued with unmounting its volume and getting rid of snapd from the system:

sudo umount /snap/core/<replace_with_number_in_your_folder>
sudo apt purge snapd

In the end i cleaned up any stuff left from snapd:

rm -rf ~/snap
sudo rm -rf /snap
sudo rm -rf /var/snap
sudo rm -rf /var/lib/snapd

In the end removing all snaps freed about 20GB of data. I will reinstall everything based on apt as usual and will then be back at maybe 15GB of saved space without snapd.

Next to this space problem all the apt installed applications start seconds faster.

Thanks for the long years of great operating system but that’s a way I won’t go with you.

sudo snap set system refresh.retain=2

which will limit the number to one live and one backup copy. But what if you also don’t need or want that backup copy? refresh.retain doesn’t allow 1:

sudo snap set system refresh.retain=1

error: cannot perform the following tasks:
- Run configure hook of "core" snap (run hook "configure": retain must be a number between 2 and 20, not "1")

I didn’t further check how to remove that but used the following one-liner to get rid of old versions:

sudo su
snap list --all | awk '$6~"disabled"{print $1" --revision "$3}' | xargs -rn3 snap remove

Next to those backups also the /var/lib/snapd/cache folder holds a massive amount of space on my machine. I considered it to be safe to delete:

sudo rm -f /var/lib/snapd/cache/*

This cleanup of snapd folder freed around 12GB of data on my root partition.

network:
  version: 2
  ethernets:
     enp6s0:
        dhcp4: true
  bridges:
     br0:
       interfaces: [enp6s0]
       dhcp4: true
       optional: false
       macaddress: <some mac>

Make sure to place a mac address in and replace the interface name enp6s0 with yours.

Then run sudo netplan apply.

The machine gets a new IP from dhcp so make sure you have it configured in your dhcp before or have a local console ;-)

This application failed to start because it could not find or load the Qt platform plugin "xcb"
in "".

Available platform plugins are: xcb.

Reinstalling the application may fix this problem.
Aborted

Trying to install xcb via apt install xcb doesn’t change the behavior. Then I found a solution by removing all libstdc++* files from the lib directory in the extracted directory:

rm lib/libstdc++*

This can be fixed by downgrading to the version before the latest one in the ubuntu repository till this bug is fixed.

Here is how I downgraded my grub installation:

sudo apt-get install grub2-common=2.02~beta2-29ubuntu0.2 \
  grub-common=2.02~beta2-29ubuntu0.2 grub-efi-amd64=2.02~beta2-29ubuntu0.2 \
  grub-efi-amd64-bin=2.02~beta2-29ubuntu0.2

sudo apt-mark hold grub-common grub-efi-amd64-bin grub2-common grub-efi-amd64

Without the hold line apt would install the latest version on the next upgrade. As soon as the bug is fixed you can remove the hold mark with

sudo apt-mark install grub-common grub-efi-amd64-bin grub2-common grub-efi-amd64

[SeatDefaults]
allow-guest=false

apt-get update
apt-get dist-upgrade
reboot

Then install docker with the following lines as root:

#!/bin/bash
apt-key adv --keyserver hkp://pgp.mit.edu:80 --recv-keys 58118E89F3A912897C070ADBF76221572C52609D
echo "deb https://apt.dockerproject.org/repo ubuntu-precise main" > /etc/apt/sources.list.d/docker.list
apt-get update
apt-get purge lxc-docker*
apt-cache policy docker-engine
apt-get install docker-engine

After this docker should be up and running. Rry using “docker ps” as root or with sudo.

sudo apt-add-repository ppa:ubuntu-audio-dev/alsa-daily
sudo apt-get update
sudo apt-get install oem-audio-hda-daily-dkms
sudo reboot