Ninety second shutdown delay: "A stop job is running for system logging service"

For the past couple of snapshots, I see this message if shutting down from run level 3:

"A stop job is running for system logging service (1 min 30 sec)"

… followed by a ninety second delay before shutdown continues.

I now also see a similar delay if shutting down from the X11 desktop, without the error message.

Is there a configuration file I can edit to eliminate or shorten the delay?

I do not see any service that includes phrase “system logging” in its description on my Tumbleweed. Find out what service it is, like

grep -r 'system logging' /usr/lib/systemd/system

I asked essentially the same question on the tde-devels list last week, and got this reply from a developer:

The 90s delays is the default timeout in systemd when somethings does not start/stop correctly… super annoying.
I usually edit the /etc/systemd/system.conf file and add these lines in:

DefaultTimeoutStartSec=5s
DefaultTimeoutStopSec=5s

This reduces the wait to 5s, which should be more than enough for any normal executing process to do what it needs.
I haven’t tried it yet, as I hadn’t run into one of these since asking the question until just minutes ago, but I’m currently trying to work out the cause ATM, which seems to be a bad NIC or motherboard going bad.

Yes… seeing that on both of my Tumbleweed machines. Introduced sometime between snapshots 20220609 and 20220614

“rsyslog” seems to be the culprit.

On my own systems I’ve temporarily removed “rsyslog” and am using “syslog-ng”, shutdown then proceeds as normal… probably warrants a bug report…

I never installed:

**erlangen:~ #** zypper se -is syslog 
Loading repository data... 
Reading installed packages... 
No matching items found. 
**erlangen:~ #** 

As rsyslog isn’t installed there is no trouble whatsoever.

Thanks for the feedback, everyone! tannington, I removed rsyslog and swapped in syslog-ng, and am now shutting down without delays. I also did the same on a Leap 15.4 VM, which had similar issues.

I have been seeing this recently and just found this thread. Seemed to be after a zypper dup but not absolutely sure. Anyway I’ve swapped to syslog-ng as well now.

Stuart

Same problem here, I opened a bug report.

Found the thread also after searching. I have this issue on my tumbleweed install.

One should carefully check whether her/his system really requires rsyslog or syslog-ng.

All my systems which do not provide Logging services for other machines and do not report to external Logging-services run happily just with journald (which comes with systemd anyway).

Regards

susejunky

I can find no reason for rsyslog to have been installed on my system, I may have missed something but I don’t think so.

Stuart

I have exactly the same problem on my laptop.
By removing rsyslog as suggested here, the problem was fixed.

However, I also have a virtual machine with virtualbox and there i have no issues.
So this is a little strange…

Many thanks for the tip.
Regards.

Are you sure that rsyslog is installed in that VM?

Regards

susejunky

I have double checked and you are right: rsyslog was not installed on that virtual machine.
But I have just installed it (just for the sake of testing) and the problem appeared on the VM also.

By removing rsyslog, the problem disappeared.

Sorry for confusion.
Again, many thanks.

Gigi

Just to point out there is bug 1200682 which is ongoing. See earlier in this thread.

Stuart