touch-packages team mailing list archive

[Didier Roche <didrocks@xxxxxxxxxx>]

Public bug reported:

One autopkgtest fails since 219-5ubuntu1. This one checks that plymouth
is started around the same time than fsckd, and that fsckd/fsck doesn't
fail if plymouth crashes.

The issue comes from debian/patches/v219-stable, which is a backport of
http://cgit.freedesktop.org/systemd/systemd-stable/log/?h=v219-stable
branch (until 16 march). The culpurit commit is
http://cgit.freedesktop.org/systemd/systemd-
stable/commit/?h=v219-stable&id=b238b0eaf71449e128480bb5a5875a4b51cafd6f.

What happens is that:
- systemd-hwdb-update.service is After=systemd-remount-fs.service (so once fsck runs)
- udev is thus started later on
- plymouth-start.service is After=systemd-udevd.service.

So fsck runs on root FS before udev is up, and so plymouth, the whole
boot sequence is delayed.

We have multiple options:
- changing the test to get that use case tested in other ways (in ubuntu, plymouth is started from the initramfs)
- reintroduce the change if we see that udev being started way later is an issue.

Attaching 2 journald boot logs, where we can see the differents time
(once with udev/plymouth starting earlier than the other one).

** Affects: systemd (Ubuntu)
     Importance: Undecided
         Status: Confirmed


** Tags: systemd-boot

** Attachment added: "normal-boot"
   https://bugs.launchpad.net/bugs/1439301/+attachment/4363164/+files/normal-boot

-- 
You received this bug notification because you are a member of Ubuntu
Touch seeded packages, which is subscribed to systemd in Ubuntu.
https://bugs.launchpad.net/bugs/1439301

Title:
  one autopkgtest failure since 219-5ubuntu1

Status in systemd package in Ubuntu:
  Confirmed

Bug description:
  One autopkgtest fails since 219-5ubuntu1. This one checks that
  plymouth is started around the same time than fsckd, and that
  fsckd/fsck doesn't fail if plymouth crashes.

  The issue comes from debian/patches/v219-stable, which is a backport
  of http://cgit.freedesktop.org/systemd/systemd-
  stable/log/?h=v219-stable branch (until 16 march). The culpurit commit
  is http://cgit.freedesktop.org/systemd/systemd-
  stable/commit/?h=v219-stable&id=b238b0eaf71449e128480bb5a5875a4b51cafd6f.

  What happens is that:
  - systemd-hwdb-update.service is After=systemd-remount-fs.service (so once fsck runs)
  - udev is thus started later on
  - plymouth-start.service is After=systemd-udevd.service.

  So fsck runs on root FS before udev is up, and so plymouth, the whole
  boot sequence is delayed.

  We have multiple options:
  - changing the test to get that use case tested in other ways (in ubuntu, plymouth is started from the initramfs)
  - reintroduce the change if we see that udev being started way later is an issue.

  Attaching 2 journald boot logs, where we can see the differents time
  (once with udev/plymouth starting earlier than the other one).

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1439301/+subscriptions