Just for reference: it was the pkill of udev, that was causing trouble in tests I did on xen guests. In that case pkill sometimes had to be redone a few times. And even then there could be even rarer cases where udev did not seem to be killable at all. But if that was not done correctly it would leave the initrd not empty and that causes (slightly overreacting there) a panic.
Just for reference: it was the pkill of udev, that was causing trouble in tests I did on xen guests. In that case pkill sometimes had to be redone a few times. And even then there could be even rarer cases where udev did not seem to be killable at all. But if that was not done correctly it would leave the initrd not empty and that causes (slightly overreacting there) a panic.