Hi Matthew, I went ahead and put together a plucky cloud image so I can try running the tests against this. I'll report back with the results.
Conceptually, the test just launches a qemu instance and the greps the logs at an interval to determine if it's either booted or hit the bug. If it's booted, the vm is killed, files cleaned up, and we start one again. If it fails, it leaves the VM running in the background so I can poke at it. I can send along the scripts, but the fiddly parts are actually setting up the machine and getting the image in a good state for testing. When I ran the Noble tests, I went through a few variations and finally settled on an initrd-less UEFI boot with the root device as a virtualized nvme device that gave the best results. I haven't tried to be this fancy with plucky yet, but did clone the uefi + nvme parts (mostly because I already had them handy). One area where I'm running into trouble with Plucky, but where I didn't with Noble, is getting the nocloud cloud-init working with the networking config. Specifically, my boot times are currently being horked by systemd-networkd-wait-online waiting for something to come online despite the fact that I've only configured loopback. If you know of a way to tell cloud-init to tell networkd that there are no more devices to wait for, I could use a hint there. In the absence of more clever ideas, I'm simply calling the test a success when it reaches the stage before the 2-minute wait, instead of waiting to see the login prompt. Best, -K -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2090972 Title: /boot intermittently fails to mount on boot To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/util-linux/+bug/2090972/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
