Bug 1112636

Summary: [Build 20181015] systemd-udevd crashed two times in openQA test on aarch64
Product: [openSUSE] openSUSE Tumbleweed Reporter: Oliver Kurz <okurz>
Component: BasesystemAssignee: systemd maintainers <systemd-maintainers>
Status: RESOLVED INVALID QA Contact: E-mail List <qa-bugs>
Severity: Major    
Priority: P5 - None CC: fbui
Version: Current   
Target Milestone: ---   
Hardware: aarch64   
OS: Other   
URL: https://openqa.opensuse.org/tests/778738/modules/gnucash/steps/21
Whiteboard:
Found By: --- Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---

Description Oliver Kurz 2018-10-20 21:05:49 UTC
## Observation

openQA test in scenario opensuse-Tumbleweed-DVD-aarch64-gnome@aarch64 fails in
[gnucash](https://openqa.opensuse.org/tests/778738/modules/gnucash/steps/21)
and I can see two crashes of systemd-udevd with core files available in https://openqa.opensuse.org/tests/778738/file/gnucash-problem_detection_logs.tar.xz. More logs are available in https://openqa.opensuse.org/tests/778738#downloads


## Reproducible

Fails since (at least) Build [20181015](https://openqa.opensuse.org/tests/776442)


## Expected result

Last good: [20181012](https://openqa.opensuse.org/tests/774724) (or more recent)


## Further details

Always latest result in this scenario: [latest](https://openqa.opensuse.org/tests/latest?test=gnome&arch=aarch64&flavor=DVD&distri=opensuse&version=Tumbleweed&machine=aarch64)
Comment 1 Franck Bui 2018-10-22 12:26:32 UTC
Well according to the logs the system is really busy to execute a process/application/whatever starving the other processes:

> Oct 20 05:27:59 linux-y5wi org.gnome.Shell.desktop[2333]: #10 0xaaab0bfe5c10 i   resource:///org/gnome/shell/ui/tweener.js:182 (0x700001cb560 @ 15)
> Oct 20 05:30:36 linux-y5wi rtkit-daemon[1609]: The canary thread is apparently starving. Taking action.
> Oct 20 05:30:36 linux-y5wi systemd[1]: systemd-udevd.service: Watchdog timeout (limit 3min)!
> Oct 20 05:30:36 linux-y5wi rtkit-daemon[1609]: Demoting known real-time threads.
> Oct 20 05:30:36 linux-y5wi systemd[1]: systemd-udevd.service: Killing process 374 (systemd-udevd) with signal SIGABRT.
> Oct 20 05:30:36 linux-y5wi rtkit-daemon[1609]: Successfully demoted thread 2344 of process 2342 (/usr/bin/pulseaudio).
> Oct 20 05:30:36 linux-y5wi rtkit-daemon[1609]: Successfully demoted thread 2343 of process 2342 (/usr/bin/pulseaudio).
> Oct 20 05:30:36 linux-y5wi rtkit-daemon[1609]: Successfully demoted thread 2342 of process 2342 (/usr/bin/pulseaudio).
> Oct 20 05:30:36 linux-y5wi rtkit-daemon[1609]: Successfully demoted thread 1958 of process 1956 (/usr/bin/pulseaudio).
> Oct 20 05:30:36 linux-y5wi rtkit-daemon[1609]: Successfully demoted thread 1957 of process 1956 (/usr/bin/pulseaudio).
> Oct 20 05:30:36 linux-y5wi rtkit-daemon[1609]: Successfully demoted thread 1956 of process 1956 (/usr/bin/pulseaudio).
> Oct 20 05:30:36 linux-y5wi rtkit-daemon[1609]: Demoted 6 threads.
> Oct 20 05:30:36 linux-y5wi chronyd[772]: Forward time jump detected!
> Oct 20 05:30:36 linux-y5wi chronyd[772]: Can't synchronise: no selectable sources

udev is one of the process that haven't been run for a while and therefore systemd reacts by killing (SIGABRT) udevd and restarts it.

IOW the coredump of udevd is the consequence of the system completely busy to run another process.

Hence closing as invalid.