systemd should sync disks, before killing all processes on shutdown

[Dimitri John Ledkov]

[Impact]

 * systemd-shutdown binary is executed to complete system shutdown. In xenial, it calls sync() after killing all processes, and not before killing them. This means that processes are racing the SIGKILL/TERM timeouts to sync all IO to disk. To mitigate this race condition, and make the shutdown process more reliable the sync() should also happen before process killing spree starts.

[Fix]
Backport upstream commit https://github.com/systemd/systemd/commit/2e79d1828a8da9b3af1b052297e3617905ec94f3

[Test Case]

 * Make sure systems still shuts down, including e.g. root on raid.

[Regression Potential]

 * Shutdowns may appear to be slower due to two sync() calls instead of one
 * However total shutdown time should not be impacted much, as there really should not be much additional IO caused by killing all processes.

[Brian Murray (brian-murray)]

Hello Dimitri, or anyone else affected,

Accepted systemd into xenial-proposed. The package will build now and be available at https://launchpad.net/ubuntu/+source/systemd/229-4ubuntu21.2 in a few hours, and then in the -proposed repository.

Please help us by testing this new package. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation on how to enable and use -proposed.Your feedback will aid us getting this update out to other Ubuntu users.

If this package fixes the bug for you, please add a comment to this bug, mentioning the version of the package you tested and change the tag from verification-needed-xenial to verification-done-xenial. If it does not fix the bug for you, please add a comment stating that, and change the tag to verification-failed-xenial. In either case, without details of your testing we will not be able to proceed.

Further information regarding the verification process can be found at https://wiki.ubuntu.com/QATeam/PerformingSRUVerification . Thank you in advance!

[Dimitri John Ledkov (xnox)]

As seen in autopkgtests, reboots are working correctly with filesystems synced to disk on shutdown, without regressions in shutdown / reboot. Thus this is sufficient to verify that the disks are still synced on shutdown.

[Launchpad Janitor (janitor)]

This bug was fixed in the package systemd - 229-4ubuntu21.2

---------------
systemd (229-4ubuntu21.2) xenial; urgency=medium

  [ Dimitri John Ledkov ]
  * udev: Mark ndb devices as inactive until connected. (LP: #696435)
  * networkd: in dhcp, change UseMTU default to true, to accept DHCP provided MTU by default.
    (LP: #1717471)
  * sysctl: apply parameters in-order, instead of randomly. (LP: #1718444)
  * networkd: apply promote_secondaries, to make DHCP lease changes work.
    (LP: #1721223)
  * shutdown: sync filesystems, before going into a killing spree.
    (LP: #1722481)
  * sysctl: do not fail, when cannot apply sysctl changes due to read-only sysfs in containers.
    (LP: #1734409)
  * networkd,wait-online: add RequiredForOnline to mark mandatory/optional links for boot.
    (LP: #1737570)

  [ David Glasser ]
  * journald: don't reduce BurstRateLimit on low disk space (LP: #1732803)

 -- Dimitri John Ledkov <email address hidden> Wed, 21 Feb 2018 13:46:37 +0000

[Łukasz Zemczak (sil2100)]

The verification of the Stable Release Update for systemd has completed successfully and the package has now been released to -updates. Subsequently, the Ubuntu Stable Release Updates Team is being unsubscribed and will not receive messages about this bug report. In the event that you encounter a regression using the package from -updates please report a new bug using ubuntu-bug and tag the bug report regression-update so we can easily find any regressions.