Bug#788050: systemd-fsck : Check disks at each reboot

[Michael Biebl]

On Wed, 7 Sep 2016 18:31:12 +0200 Michal Schmidt <mschmidt at redhat.com>
wrote:
> systemd-fsckd's event loop terminates if nothing happens for 30 seconds
> (IDLE_TIME_SECONDS). Usually fsck writes progress updates more frequently
> than that, but the interval is not guaranteed. So systemd-fsckd may exit
> by itself while fsck is busy working. When this happens, fsck will
> receive SIGPIPE as soon as it tries to write the next progress update.
> 
> It is possible to reproduce this using the dm-delay DM target for
> a block device.
> 
> The bug is specific to Debian/Ubuntu, because systemd-fsckd is not
> present in upstream.

So, with Michal's hint I've setup a really slow dm-delayed device. Then
I ran a systemd-fsck@ check on that device.

systemd-fsckd does indeed exit before the fsck process has finished,
*but* I don't get a failing fsck because of that.

So I wonder if there were some changes in either util-linux or e2fsprogs
which handle SIGPIPE differently now in fsck.
Maybe tytso has some insight here.

For anyone who was able to reproduce the issue, which versions of
util-linux and e2fsprogs do you have installed which show this problem?
Can you still reproduce it on an up-to-date stretch system?

Regards,
Michael


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

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: OpenPGP digital signature
URL: <http://alioth-lists.debian.net/pipermail/pkg-systemd-maintainers/attachments/20160910/c5287c78/attachment-0002.sig>