[systemd-devel] How to debug stop job running into time-out?

[Lennart Poettering]

On Mi, 24.10.18 17:07, Paul Menzel (pmenzel+systemd-devel at molgen.mpg.de) wrote:

> Dear systemd folks,
> 
> 
> With Debian Sid/unstable (systemd 239) and GDM 3.30.1, sometimes
> powering off the system, it takes two minutes and the message
> below is shown on the screen.
> 
> > A stop job is running for User Manager for UID 1000 (1min 3s / 1 min 59s)
> 
> What ways are there to debug this issue. The problem is, it’s
> not always reproducible.
> 
> Is there a key combination for dumping the necessary information?
> Like what process(?) cannot be stopped?

Well, it's the service in question which might be hanging. And in this
case that's the "systemd --user" instance of UID 1000. Most likely
this means some user service doesn't want to shut down, which doesn't
let the systemd --user instance shut down, and hence not the systemd
--system instance either...

Lennart

-- 
Lennart Poettering, Red Hat