I now have Fedora 34 running on all 7 nodes of my Turing Pi 1 cluster. Tedious to install, so these are just my notes.
Compile https://github.com/raspberrypi/usbboot
Insert the compute module to program, set the jumper to flash mode, connect the cable, power it on and run ./rpiboot
/dev/sdX will appear once the CM is in the right mode. Following the instructions here:
arm-image-installer --image=$HOME/Fedora-Server-34-1.2.aarch64.raw.xz --target=rpi3 --media=/dev/sdb --resizefs --relabel --addkey $HOME/.ssh/id_rsa.pub
This takes quite a long time to run. You also need to kill initial-setup:
virt-customize -a /dev/sdb --link /dev/null:/etc/systemd/system/initial-setup.service --selinux-relabel
If it’s successful, reboot the Turing Pi with the jumper in the boot setting and check the first CM comes up (I used an HDMI monitor)
Repeat with the other 6 boards(!)
Edit: I couldn’t work out how to get stable MAC addresses. To get a stable MAC address:
nmcli c modify "Wired connection 1" 802-3-ethernet.cloned-mac-address aa:bb:cc:dd:ee:ff
Richard WM Jones 