Bug 20688 - console-kit-daemon found consuming 96-100% of CPU
Summary: console-kit-daemon found consuming 96-100% of CPU
Status: RESOLVED WORKSFORME
Alias: None
Product: ConsoleKit
Classification: Unclassified
Component: Daemon (show other bugs)
Version: unspecified
Hardware: x86 (IA32) Linux (All)
: medium normal
Assignee: william.jon.mccann
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-03-16 06:59 UTC by Whit Blauvelt
Modified: 2009-09-23 19:27 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments

Description Whit Blauvelt 2009-03-16 06:59:18 UTC
Running a current Ubuntu Intrepid on a VIA C7-D-processor-based server, which normally runs with a load of < 0.10, the system for over a day went to a load of 1 and over. Checking with top showed that console-kit-daemon was consuming 96-100% of CPU. Stopping it with killall fixed the problem. Of course it respawns. The problem hasn't recurred, but I have a cron job doing a killall of console-kit-daemon once a day to be sure. This is with the consolekit-0.2.10-1ubuntu deb.

Can this be anything other than a memory leak? The system had been up for 100 days when the problem showed up - can't say precisely when it started. There was very little console activity on it during that time - an xfce (Xubuntu) screen with a terminal session open on it, and typical access being via ssh. So it's not like console-kit-daemon had a lot to be active with.
Comment 1 Whit Blauvelt 2009-03-16 07:05:50 UTC
Edit: That's a VIA Esther in that box, not a C7-D.
Comment 2 william.jon.mccann 2009-07-29 17:18:04 UTC
Can you try to profile it to see what is going on?  Try attaching gdb to it.
Comment 3 Whit Blauvelt 2009-07-29 18:50:41 UTC
Don't know if it's repeated. My systems are now at consolekit 0.3.0-2ubuntu4. On the one that had the problem, there's still a cron job to kill console-kit-daemon overnight, but haven't tracked whether it got to the runaway state again. On a second VIA box similarly configured (but with the C7-D cpu rather than Esther) with no such automated kill cycle over these months it has never run away. Not sure if the condition can be recreated.

In any case about to take a vacation, so can't test until after next week. Maybe I'll set up something to watch its resource consumption on the box it once had the problem on, take the kill out of cron, and see if the situation replicates.
Comment 4 william.jon.mccann 2009-09-23 19:27:43 UTC
Haven't had any other reports of this.  Please reopen if you can reproduce and get a backtrace.  Thanks.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.