Better write-up is coming. Until then, use the source:
- Add the ppa for it:
sudo add-apt-repository ppa:ubuntu-raspi2/ppa
- Edit /etc/apt/sources.list.d/ubuntu-raspi2-ubuntu-ppa-eoan.list. Change eoan to bionic, because bionic is the latest supported release.
sudo apt update
sudo apt install libraspberrypi-bin
This will install
This can be easily automated using ansible:
tasks: - name: Set up the ppa for vcgencmd become: true apt_repository: repo: ppa:ubuntu-raspi2/ppa codename: bionic - name: Install ubuntu raspberry pi support library apt: name: libraspberrypi-bin
Taken from here.
Download the Ubuntu raspberry pi image from the Ubuntu Server for Raspberry Pi page.
Then use the
gnome-disks program to copy it onto an SD card.
And then wait.
Once the image has been written, remove the SD card and then reinsert it (if necessary; about half the time it just mounted the new partitions for me.) Then you need to do things:
- Create an empty file called “ssh” on the system-boot partition. There are any number of ways to do this, but something like
touch /media/$USER/system-boot/sshshould work if you like working from the command line.
- There will be a file in the system-boot partition named
nobtcmd.txt. This file contains the kernel command line options used when linux is started. To work with k3s, you want to add
cgroup_memory=1 cgroup_enable=memoryto the end of the line.
- If you want to overclock your Pi, you should add your overclock parameters to
usrcfg.txtin the system-boot partition. This is slightly different from the way you do it in Raspbian, where you edit
config.txtdirectly; The Ubuntu boot system has things broken out a bit more.
You need to do these things for each SD card you’re setting up. Once this is done, you will have a system image that will boot Ubuntu on your cards, and you’re ready to start deploying things.