From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.7.6) Gecko/20050223 Firefox/1.0.1 Description of problem: Version-Release number of selected component (if applicable): kdelibs-3.3.1-2.6.FC3 yum-2.2.0-0.fc3 up2date-4.3.47-5 How reproducible: Always Steps to Reproduce: 1. launch up2date (gui) or execute 'yum update' 2. install updates (there is no problem when there are no updates available) Actual Results: A kdeinit proces is started as user root and uses 100% cpu time untill it is manually terminated (a simple kill <pid> does the trick). Killing the kdeinit proces doesn't seem to affect anything. There are no errors. Expected Results: System should return to idle after the updates are installed. In fact I don't understand why a kdeinit proces is started as user root in the first place. Additional info: This problem has been there for a while. Don't recall any updates to yum or up2date but updates to kde are installed since. System is vanilla fc3 + fedora extras with a few exceptions: xmms, xawtv, kleds and gnome-cpufreq-applet from rpms, kcpuspeed from source and gftp rebuild rpm. Don't think this has anything to do with this problem, just trying to provide as much info as possible.
Created attachment 112065 [details] lsof output for kdeinit pid Don understand why lot's of files in /usr/share/applnk are open.
I'm seeing this also. I have two kdeinit process taking 98% of their respective CPU's (two-processor SMP box). One is running a root, the other as my regular user account. The 'ps' utility indicates that both of these kdeinit's are running the kded daemon. I'm not sure when it started so I can't directly pin this on applying an update. I did an strace on both processes. Both are in a tight loop making this set of system calls: gettimeofday({1111122908, 713705}, NULL) = 0 select(17, [3 4 5 7 9 10 13 16], [], [], {0, 682}) = 1 (in [9], left {0, 1000}) gettimeofday({1111122908, 714744}, NULL) = 0 gettimeofday({1111122908, 714926}, NULL) = 0 gettimeofday({1111122908, 715110}, NULL) = 0 gettimeofday({1111122908, 715364}, NULL) = 0 ioctl(4, FIONREAD, [0]) I'm also seeing lots of open files by the regular-user instance of the runaway kdeinit/kded process. The root runaway instance has just a few files open under it's own /proc/[pid] directory. Same version of kdelibs as the reporter.
kdeinit is running kded here also. Strace looks like this: select(11, [3 4 5 7 9 10], [], [], {0, 897099}) = 1 (in [9], left {0, 898000}) ioctl(4, FIONREAD, [0]) = 0 select(11, [3 4 5 7 9 10], [], [], {0, 896826}) = 1 (in [9], left {0, 897000}) ioctl(4, FIONREAD, [0]) = 0 select(11, [3 4 5 7 9 10], [], [], {0, 896569}) = 1 (in [9], left {0, 897000}) ioctl(4, FIONREAD, [0]) = 0 select(11, [3 4 5 7 9 10], [], [], {0, 896312}) = 1 (in [9], left {0, 897000}) ioctl(4, FIONREAD, [0]) = 0 select(11, [3 4 5 7 9 10], [], [], {0, 896055}) = 1 (in [9], left {0, 897000}) ioctl(4, FIONREAD, [0]) = 0 select(11, [3 4 5 7 9 10], [], [], {0, 895780}) = 1 (in [9], left {0, 896000}) ioctl(4, FIONREAD, [0]) = 0 select(11, [3 4 5 7 9 10], [], [], {0, 895523}) = 1 (in [9], left {0, 896000}) ioctl(4, FIONREAD, [0]) An endless loop op select and ioctl calls, but the numbers for select keep changing.
This comment is just a test (see bug #151461).
Fedora Core 3 is now maintained by the Fedora Legacy project for security updates only. If this problem is a security issue, please reopen and reassign to the Fedora Legacy product. If it is not a security issue and hasn't been resolved in the current FC5 updates or in the FC6 test release, reopen and change the version to match. Thank you!
Fixed in later releases