xorg eat 100% of the CPU when I use superkaramba and lock my desktop for a while ([apport] Xorg crashed with signal 5)

[totya]

Hi,

I have to kill the X server because eat 100% CPU. I can reproduce the problem on two of my machine.

If I use superkaramba on my desktop and I lock the screen, after 10-15 minutes or 1 hour the X process will be going to eat my CPU.

I'm going to attach more logs to this bugreport later.

ProblemType: Crash
Architecture: i386
Date: Tue Apr 24 07:12:01 2007
DistroRelease: Ubuntu 7.04
ExecutablePath: /usr/bin/Xorg
Package: xserver-xorg-core 2:1.2.0-3ubuntu8
PackageArchitecture: i386
ProcCmdline: /usr/bin/X -br -dpi 120 :0 vt7 -auth /var/run/xauth/A:0-0bDlcl
ProcCwd: /etc/X11
ProcEnviron: PATH=/bin:/usr/bin:/sbin:/usr/sbin
Signal: 5
SourcePackage: xorg-server
Stacktrace: #0 0x0809c309 in ?? ()
StacktraceTop: ?? ()
ThreadStacktrace:

Uname: Linux algol 2.6.20-15-generic #2 SMP Sun Apr 15 07:36:31 UTC 2007 i686 GNU/Linux
UserGroups:

[totya (totya)]

Hi,

Here's some update...

I was who send the kill -5 signal to the X server, because I would like to see a crash report and I would like to help to find the root of the problem. My machine has a NVIDIA 7600 card. I use the nvidia driver from the ubuntu repository. However I can reproduce this symptom on my notebook too, which has an ATI FireGL card with the fglrx driver. So I think this isn't nvidia or ati specific problem. Somehow the superkaramba and the kscreensaver application do something and the xorg is going to crazy. But this is my theory only. However I did not see this problem if I not use superkaramba on my desktop.

I'm going to attach my xorg.conf file.

If you need more information please do an update!

[totya (totya)]

I'm going to set this confirmed, because I'm able to reproduce this problem on another machine (on different hardware environment).

[totya (totya)]

I uploaded some file.

I uploaded my package list.
I uploaded my xorg.conf file.
I uploaded 2 trace file which created the following way: strace -p PID of X -o trace file
I uploaded a process list which created in the time of the second trace file.

I hope these files contains some useful information for you.

[totya (totya)]

Hi,

I added kubuntu-team as a subscripted one, because I think this problem is superkaramba related. I also found a bug report on bugs.kde.org:

http://bugs.kde.org/show_bug.cgi?id=143255

[totya (totya)]

I think this problem is superkaramba related instead of xorg-xserver. However I sent the kill signal to the xorg process, so apport sent this report as an xorg-server error to the launchpad.

[totya (totya)]

I think this problem is superkaramba related.

[footer (footer)]

I can confirm this problem on Feisty Fawn 7.04, KDE 3.5.6 and Superkaramba 0.41. I have two Superkaramba applets running and after sitting idle for awhile, my desktop will freeze for 20-30 seconds after which I am able to use it normally again (no hard reset thank goodness!). Very annoying problem!

Thanks!

[nukedathlonman (areginato)]

Strange, but I can add further verification of this bug... With superkaramba loaded, Xorg would ramp up to 100% usage over 30 minutes when my laptop was left to idle. 15 minutes after it reach top speed it the system would crash if I did anything except kill Xorg. I hadn't experienced this before upgrading to Feisty, and was experiencing it thereafter. I was sure I must have had something mis-configured (I had made a lot of video and screen related changes to play with dual head operation, my HDTV just prior to the upgrade). After endless debugging I gave up trying to solve the issue, and reported it under bug 51991 because it resembled everything I could debug. Even top wasn't showing superkaramba as eating up the CPU cycles, just Xorg. After reading this bug report, I decided to try unloading superkarama, and low and behold, the problem has completely disappeared. Thanks for the reporting, I never would have thought of this - hope this bug can be fixed!!

[ZdadrDeM (zdadr-dem)]

Same here. After unloading superkaramba the problem is gone.
There will be also no problem if only one superkaramba applet is running.

[Borahshadow (codyregister)]

I think I have the same problem
I'm running feisty and I never saw anything like this under edgy
I have 2 superkaramba applets open and when I'm using my computer xorg eats about 10-15% of my cpu and I've noticed when I leave my laptop on overnight that something eats enough cpu to scale up my processor to full speed and make my fan start spinning full speed
sometimes after my screen has gone to screen saver(actually after it has turned off the screen) and I come and wiggle the mouse it shows my desktop but I can't do anything(could be due to xorg eating 100% not sure though) I have to ctrl-alt-backspace to do anything
top shows that xorg right now is using 10-11% and superkaramba is using 5-10%

[footer (footer)]

I think I identified incorrectly what version of Superkaramba I was using in my post above (2007-06-08). I said .41 but at the Superkaramba web site, it's only up to .39 (as of today anyway!). I tried installing .39 from source but keep running into this problem:

checking for unsetenv... yes
checking if unsetenv needs custom prototype... no
checking whether byte ordering is bigendian... no
checking for MAXPATHLEN... 4096
checking for Python directory... /usr/local
checking for Python2.4... header no library no modules /usr/lib/python2.4
checking for Python2.3... header no library no modules no
checking for Python2.2... header no library no modules no
checking for Python2.1... header no library no modules no
checking for Python2.0... header no library no modules no
checking for Python1.5... header no library no modules no
checking for libxmms... ./configure: line 43081: xmms-config: command not found
./configure: line 43083: xmms-config: command not found
no
checking for main in -lknewstuff... yes
checking for main in -lkvm... no
checking if doc should be compiled... yes
checking if superkaramba should be compiled... no
configure: creating ./config.status
fast creating Makefile
fast creating doc/Makefile
fast creating doc/superkaramba/Makefile
fast creating superkaramba/Makefile
fast creating superkaramba/doc/Makefile
fast creating superkaramba/icons/Makefile
fast creating superkaramba/mimetypes/Makefile
fast creating superkaramba/src/Makefile
config.pl: fast created 8 file(s).
config.status: creating config.h
config.status: config.h is unchanged
config.status: executing depfiles commands

Superkaramba can't be compiled
because of missing Python libraries/headers.

I'm running Kubuntu Feisty 7.04 64-bit so not sure what's going on. As I understand it, Superkaramba is part of KDE in Feisty. Here are the details from Adept:

superkaramba: installed
     Section: universe/kde
     Installed Size: 1544K
     Maintainer: Ubuntu Core Developers
     Candidate Version: 4:3.5.6-0ubuntu2
     Installed Version: 4:3.5.6-0ubuntu2

Has anyone with a 64-bit version of Feisty successfully installed Superkaramba 0.39 and if so, does this problem go away?

Thanks!

[Brett A. Taylor (realbt)]

Bump. This is affecting me too. I also thought it was bug 51991.

I can provide additional tracing, etc. as needed.

[Patrick Salami (pat-entitycom)]

this is apparently similar to bug 51991 as well as bug 120347.

I have exactly the same problem with xorg taking 100% cpu if the computer is left for a couple of hours. The intervals are random though, sometimes it happens after only a few mins.
I haven't tried running it without superkaramba yet, but I might try that if no other fix is suggested.

[Sitsofe Wheeler (sitsofe)]

This is also similar to another superkaramba issue - Bug #103603 .

[Bryce Harrington (bryce)]

Some have reported that even after disabling superkaramba, the issue is still there, just much less frequent. So I think we have two bugs - one specific to superkaramba (this bug 109507) that somehow something it is doing is exacerbating the second bug, which is specific to xorg (bug 51991) and causes the increased cpu usage.

Since we already have bug 51991 for tracking the xorg-specific aspect of this problem, I'm closing the xorg task in this (sk-specific) bug.

[tripmix (tripmix)]

I have this too. I'm on debian sid amd64 arch.

[tripmix (tripmix)]

I just noticed this is from 2007, thats a bit odd as it just started happening to me about a month ago and I never noticed anything before... Iv been using the same setup for quite a few years and allways keep my system up to date. Only changes Iv made to hardware is a newer nvidia graphics card and a new MB.

[replica9000 (replica9000)]

I also had this problem on my laptop, (Debian Sid) takes about an hour of being idle before the cpu usage would get to the point where even killing xorg would save me from a reboot. Top would show that xorg was using all the CPU usage, but found it was SuperKaramba causing the problem. Seems to happen when I use it to monitor my hard drive usage and/or memory usage. I made a work around for this problem by creating a small script that would kill and restart SuperKaramba, and then using KAlarm to run this script every 45 minutes.

The script:

#! /bin/bash

killall superkaramba
sleep 10s
superkaramba /home/replica/systeminfo.theme

I'm running the same OS on my desktop, and this problem doesn't happen there.

[Jonathan Thomas (echidnaman)]

Is this still a problem in Intrepid?

[totya (totya)]

> Is this still a problem in Intrepid?

I've left Kubuntu and I'm using Ubuntu Intrepid nowadays. So unfortunately I'm unable to answer your question.

[footer (footer)]

I've been using Kubuntu Intrepid since it came out in late October (2008). Although I'm only running 1 Superkaramba applet, this problem has not occurred on my system.

[Jonathan Thomas (echidnaman)]

Ok, I think that's probably enough to close this for now. Thanks for testing.