Bug#931267: times out and drops into useless emergency shell with fsck still ongoing
Luca Boccassi
bluca at debian.org
Mon May 27 17:24:56 BST 2024
Control: close -1 256~rc3-1
On Mon, 1 Jul 2019 14:46:09 +0100 Steve McIntyre <steve at einval.com>
wrote:
> Control: severity -1 normal
>
> Hey Michael,
>
> Agreed on the downgrade...
>
> On Mon, Jul 01, 2019 at 02:14:26PM +0200, Michael Biebl wrote:
> >
> >with your additional information about the faulty fstab entry,
> >I had another look.
> >a/
> >I first tried with a non-existing device. I also made sure the
> >concurrently running fsck takes longer then 90s.
> >
> >The result is:
> >https://people.debian.org/~biebl/bug931267/boot-missing-device.mp4
> >
> >systemd will indeed start the emergency shell after 90s, although
the
> >fsckd process is still ongoing and clobbers the output of the login
prompt.
> >
> >Once the fsck is done, simply hitting enter one can log in without
problems.
> >
> >b/
> >Next I tried with a faulty mount point where the device exists but
the
> >mount options are non-sense, so will trigger a mount failure.
> >The emergency shell is immediately started while fsck is still
ongoing.
> >See
> >https://people.debian.org/~biebl/bug931267/boot-failing-mount.mp4
> >
> >I guess this is basically what happened in your case.
> >The login prompt was again still usable.
> >
> >Given this, I'm inclined to downgrade the severity.
> >
> >I'm not entirely sure how to fix this though.
> >Should systemd delay the start of the sulogin prompt until all fsck
> >processes have finished? You can't interact with the system for a
> >potentially very long time (the same way basically as is the case
now).
> >The only thing you'd gain is that the login prompt in such a case
> >doesn't look clobbered.
>
> Right. Starting things on a busy console like now is confusing for
> users. I'm not sure there *is* a good answer here, tbh. :-/
>
> Maybe(?) it would be possible to steal a few more characters of each
> line of the terminal output at boot for a tiny status message? The
you
> could have that show that the boot has hit errors? Similar to the
> existing LSB-style [ OK ] or [FAILED] messages for each servive, but
> to show overall system status?
>
> My own system looked a lot messier than what you're showing in your
> simple video, of course - other services starting up around this
added
> a lot of noise. My server runs lots of services. Of course, I didn't
> get to capture the output directly, just the logs.
fsckd has now been dropped (after being out of tree for many years), so
this won't apply anymore, closing.
--
Kind regards,
Luca Boccassi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: This is a digitally signed message part
URL: <http://alioth-lists.debian.net/pipermail/pkg-systemd-maintainers/attachments/20240527/21ab0129/attachment-0001.sig>
More information about the Pkg-systemd-maintainers
mailing list