FS#23955 - [kded4] cpu load 100% after restart X

Attached to Project: Arch Linux
Opened by synflag (synflag) - Wednesday, 27 April 2011, 07:51 GMT
Last edited by Andrea Scarpino (BaSh) - Wednesday, 19 October 2011, 09:04 GMT
Task Type Bug Report
Category Upstream Bugs
Status Closed
Assigned To Ronald van Haren (pressh)
Andrea Scarpino (BaSh)
Architecture All
Severity High
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 3
Private No

Details

Description:

After restart X with ctrl+alt+backspace and wait a few minutes, kded4 overload the system.

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1465 synflag 20 0 629m 29m 15m S 100 0.4 4:53.91 kded4
1881 synflag 20 0 612m 28m 14m S 100 0.4 4:53.88 kded4


And the bug fix only restart the machine.
If start in runlevel 3 with startx, when ctrl+alt+backspace is used, i can see the child no killed of kdeinit, and pressing enter and killing the kded4 and kdeinit the cpu load is normal, again startx and all fine.
I read in KDE forum that is a problem in soprano, but, i install the last version of soprano of testing and dependencies and the bug persist.
Additional info:
* package version(s)

KDE 4.6.2
kdelibs 4.7.2-6
* config and/or log files etc.


Steps to reproduce:

start the system, restart X and wait few minutes, open firefox, etc, kded4 up to 100% of cpu load.
This task depends upon

Closed by  Andrea Scarpino (BaSh)
Wednesday, 19 October 2011, 09:04 GMT
Reason for closing:  Upstream
Additional comments about closing:  Please follow the upstream bug report
Comment by synflag (synflag) - Wednesday, 27 April 2011, 08:01 GMT
Fix of KDE Team:

http://bugs.kde.org/show_bug.cgi?id=159027

plain txt patch: http://bugsfiles.kde.org/attachment.cgi?id=25982

Note: i not know how apply this patch
Comment by Greg (dolby) - Wednesday, 27 April 2011, 09:12 GMT
This patch is from 2008. If its still not applied to strigi i would recommend working with upstream in order to apply it or solve the bug differently. Arch doesnt apply patches the upstream developers dont.
Comment by synflag (synflag) - Wednesday, 27 April 2011, 09:34 GMT
If developers do not apply patches, then there is nothing to do, 99% of the bugs are for things outside of Arch, KDE, pulseaudio, alsa, skype, kdelibs, then you do not, I thought that if they could do so.
Thanks Greg
Comment by Greg (dolby) - Wednesday, 27 April 2011, 10:06 GMT
Since the patch on the KDE bug report comes from a KDE developer, AFAIK the one more relevant than anyone to this semantic stuff, id query him about this. At the very least you could comment on the KDE bug report.
Comment by synflag (synflag) - Thursday, 28 April 2011, 12:03 GMT
Confirmed more users with the same bug:

https://bbs.archlinux.org/viewtopic.php?id=116355&p=3

Last edit: 2011-04-25 01:39:01

Previously the same bug related with ntrack version closed and fixed: https://bugs.archlinux.org/task/23612
This bug is no related with ntrack in this case, the bug appear when restart X, the same problem that the forum https://bbs.archlinux.org/viewtopic.php?id=116355&p=1

For the moment start session in runlevel 3, and add kdekillall -9 kded4 in the last line of /usr/bin/startkde, and work, kill all kded4 loops.

Some idea?, wait for fix, or new kde version, or whatever.
Comment by Jelle van der Waa (jelly) - Thursday, 05 May 2011, 10:11 GMT
You should talk to upstream about the problem, and why the patch didn't get applied. We can't do anything.
Comment by synflag (synflag) - Saturday, 07 May 2011, 06:22 GMT
After new update of kdelibs and kde to 4.6.3 kded4 no overload the CPU anymore, only fork an kded4 [kdeinit] after X restart, if restart 4 times, 4 kdeinit4: kded4 [kdeinit] appear in process, so, 8 restart of X, then 8 processes appear.
Comment by Instor Lesser (wingrime) - Wednesday, 25 May 2011, 19:06 GMT
Now I have this problem, at fresh home dir and after updates
Comment by Andrea Scarpino (BaSh) - Wednesday, 25 May 2011, 19:06 GMT
@Instor
I don't think is a packaging bug, please report this upstream
Comment by Jim Yang (yangjeep) - Friday, 27 May 2011, 10:04 GMT
Same behavior on my laptop.
Only one kded4 in my "top" command.
Comment by Andrea Scarpino (BaSh) - Monday, 30 May 2011, 08:10 GMT
please test the system with ntrack 1:13-2
Comment by Jim Yang (yangjeep) - Monday, 30 May 2011, 14:24 GMT
Attached the "ps -aux" sorted by cpu occupancy. Here the 60.7 refers to the 60.7% of two cores in total.

yangjeep 2897 0.3 1.1 244764 34104 ? Sl 22:13 0:00 kdeinit4: krunner [kdeinit]
yangjeep 2796 0.3 0.7 173436 23072 ? Sl 22:13 0:00 kdeinit4: kded4 [kdeinit]
yangjeep 3072 0.8 0.6 133704 20456 ? Sl 22:13 0:02 kdeinit4: konsole [kdeinit]
yangjeep 3121 0.8 0.9 125704 28528 ? SNl 22:13 0:02 /usr/bin/nepomukservicestub nepomukstrigiservice
yangjeep 2816 0.9 1.1 157464 35672 ? Sl 22:13 0:02 kwin -session 10275247240249000130625496400000014190000_1306750443_652242
yangjeep 2906 1.1 0.6 196728 19672 ? SNl 22:13 0:03 /usr/bin/nepomukservicestub nepomukstorage
yangjeep 2828 2.2 2.7 363308 85544 ? Sl 22:13 0:06 kdeinit4: plasma-desktop [kdeinit]
yangjeep 2235 60.7 0.7 173368 23000 ? Sl 22:09 5:04 kdeinit4: kded4 [kdeinit]
   kded4.log (7.8 KiB)
Comment by Jim Yang (yangjeep) - Monday, 30 May 2011, 14:34 GMT
I think there is no point to test the system with ntrack.
Comment by Andrea Scarpino (BaSh) - Monday, 30 May 2011, 14:37 GMT
you are wrong: https://bugs.launchpad.net/ntrack/+bug/755608
https://bugs.archlinux.org/task/24370#comment77771
please update your system and see if this is fixed.
Comment by Jim Yang (yangjeep) - Monday, 30 May 2011, 14:59 GMT
No use...
I notice that I've already installed ntrack 1:13-2
and my kde4 is 4.6.3-1

But this problem is still there.
Comment by Jim Yang (yangjeep) - Tuesday, 31 May 2011, 09:50 GMT
I believe this problem is related to kded4 and networkmanager.
once i remove the networkmanager from DAEMON in rc.conf this problem no longer exists.
It's inherited from previous version of kde.
Comment by Andrea Scarpino (BaSh) - Monday, 18 July 2011, 07:44 GMT
Status?
Comment by Jim Yang (yangjeep) - Wednesday, 20 July 2011, 09:45 GMT
still unresolved in current version
Comment by Andrea Scarpino (BaSh) - Tuesday, 02 August 2011, 08:58 GMT
Status? Does upstream know about this?
Comment by Jim Yang (yangjeep) - Wednesday, 03 August 2011, 02:33 GMT
the bug is already in upstream.
https://bugs.kde.org/show_bug.cgi?id=217364

Loading...