I (accidentally) verified that without CPU limit this modified root disk would boot fine.
And will now run it with 40% over lunchtime (a bit less to more surely trigger it - and since I don't "wait" for it to complete lower % won't hurt).
But that still ran into the issues.
At least I have an emergency shell there interactively now that I can take a look before the next tries.
Also not only disk fails but also (probably dependency) Timed out waiting for device dev-ttyS0.device.
Running just with tgt-boot-test in interactive mode, no user-data and no extra disk.
Issue still reproducible with that.
Saving the image to try different modifications while being able to get back to the current state.
Then I started where Scott already experimented. The timeouts.
I was trying the big hammer first setting all of them.
# Inside: callback --system-mounts disk.img /bin/bash
sudo mount-image-
/etc/fstab: rootfs / ext4 defaults, x-systemd. device- timeout= 600,x-systemd. mount-timeout= 600,_netdev 0 0
LABEL=cloudimg-
$ mkdir etc/systemd/ user.conf. d tartSec= 600s > etc/systemd/ user.conf. d/longtimeout. conf
$ echo DefaultTimeoutS
I (accidentally) verified that without CPU limit this modified root disk would boot fine.
And will now run it with 40% over lunchtime (a bit less to more surely trigger it - and since I don't "wait" for it to complete lower % won't hurt).
But that still ran into the issues.
At least I have an emergency shell there interactively now that I can take a look before the next tries.
Also not only disk fails but also (probably dependency) Timed out waiting for device dev-ttyS0.device.