[systemd-bugs] [Bug 70593] random 90 sec delay on shutdown/reboot (A stop job is running for User Manager for 0)

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=70593

--- Comment #37 from Simon McVittie <simon.mcvittie at collabora.co.uk> ---
Everyone who has reopened this bug/reported it as happening recently: you have
seen the same *symptoms*. That does not necessarily mean that you have all
encountered the same *bug* as the original reporter, or even the same bug as
each other.

(In reply to Mike C from comment #28)
> Feb 18 20:14:53 home1 systemd[589]: dbus.service: State 'stop-final-sigterm'
> timed out. Killing.

This means that either dbus-daemon itself, or a child process that was launched
by D-Bus activation, didn't terminate promptly when it received the TERM
signal. IMO it is much more likely to be a different child process (some random
D-Bus session service), because dbus-daemon's own SIGTERM handling is rather
simple.

If you enable the systemd debug shell and use it to run "systemd-cgls" during
this 90 second wait, you'll probably see that all processes in the dbus.service
cgroup below user at 1000.service were terminated by systemd sending the TERM
signal, except for some daemon (for example it might be kglobalaccel5 in
Comment #36) or a small number of daemons. If so, the cause of your slow
shutdown is very likely to be a bug in those daemons - they do not die when
sent the TERM signal - similar to the kglobalaccel5 bug described in Comment
#36.

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/systemd-bugs/attachments/20160411/a69c6f21/attachment.html>