Bug#918322: initramfs-tools: kernel fails to boot with "Gave up waiting for root file system device"

[Isaac Gelado]

On Tue, Jan 8, 2019 at 8:38 AM Michael Biebl <biebl at debian.org> wrote:
>
> Am 08.01.19 um 16:43 schrieb Isaac Gelado:
> > Adding sleep 1 does not work. I also check that the -v flag is only
> > needed in one of the calls to udevadm.
>
> Urgh, this is getting really weird.
> -v should not affect the behaviour of udevadm, but only generate
> additional output which might slow it down a bit.
> So I was suspecting a race between the startup of systemd-udevd and the
> first "udevadm trigger" call and you avoided that race by slowing down
> udevadm.
> Apparently this doesn't seem to be the case if an explicit "sleep 1"
> does not help. I assume increasing that from 1 to say 5 seconds doesn't
> make a difference?

Definitively the `-v` option delays udevadm by less than one second,
so I would assume a `sleep 5` is not going make any difference...
Anyways, I can try that out later today (this only happens on my work
desktop and every reboot requires saving quite a few screen sessions).

Could the race be in the interaction between udevadm and the udev daemon?

> (And I assume you made sure to rebuild the initramfs
> after modifying /usr/share/initramfs-tools/scripts/init-top/udev)

Yes, otherwise the system would have boot without errors and lots of output.

As a side hint that might be helpful. I had two kernels 4.18 and 4.19
in my system. This problem started when I first boot 4.19, while I was
able to boot 4.18 without problems. At some point I decided to do an
`update-initramfs -k all -u` and the problem started to also happen
when booting 4.18. I do not understand why when upgrading udev/systemd
from 239-15 to 240-2 the trigger only updated the iniramfs for 4.19.


>
>
>
>
> --
> Why is it that all of the instruments seeking intelligent life in the
> universe are pointed away from Earth?
>