trackerd 100% cpu

[Keepiru]

Binary package hint: tracker

trackerd is using 100% CPU. This never happened before I upgraded to 9.04. I'm filing a separate bug from #239391 because I do not have any log messages, and so it may be a different cause.

Here's some strace:

gettimeofday({1240752887, 960549}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=9, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}], 5, 0) = 0 (Timeout)
gettimeofday({1240752887, 960608}, NULL) = 0
poll([{fd=4, events=POLLIN}, {fd=3, events=POLLIN}, {fd=9, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7, events=POLLIN}], 5, 0) = 0 (Timeout)
gettimeofday({1240752887, 960666}, NULL) = 0

I still have it running. Let me know if there's any other data you'd like me to collect.

ProblemType: Bug
Architecture: i386
DistroRelease: Ubuntu 9.04
NonfreeKernelModules: nvidia
Package: tracker 0.6.93-0ubuntu1
ProcEnviron:
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: tracker
Uname: Linux 2.6.28-11-generic i686

[Chris Coulson (chrisccoulson)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. Please answer these questions:

* Is this reproducible?
* If so, what specific steps should we take to recreate this bug?

This will help us to find and resolve the problem.

[Keepiru (keepiru)]

Unfortunately, it's sporadic. It's reproducible by logging in, and doing normal everyday things for a while, and eventually it gets stuck there. I've not noticed a specific sequence of events that triggers it.

It also reindexes every time I log in. I'm not sure if that's related. I kind of suspect that it sometimes gets stuck on some file when indexing, but sometimes it finishes and everything works fine for a while.

Sorry to not have a simple test case. :)

[Joost Cassee (jcassee)]

Just upgraded to Karmic. Still affected by this bug. Unfortunately, it just happens after a while.

[Jerry Chong (zanglang)]

Is this bug related to #458995?

[Joost Cassee (jcassee)]

@Jerry I'm not sure, I don't have any ignored files, and my tracker log files are empty.

[mchibani (mchibani)]

It seems that this is still around. Any workaround ? Does this happen in 9.10 ?

--------------------
Ubuntu 9.04
Compaq 6710b
Intel(R) Core(TM)2 Duo CPU T7300 @ 2.00GHz

[Keepiru (keepiru)]

mchibani @#7: Joost @#4 says it affects Karmic 9.10. My current workaround is to switch to beagle.

[Joost Cassee (jcassee)]

Yup, sorry to say it's still happening in Karmic.

[Raf (4263004-noduck)]

I seem to have the same problem on Lucid (0.6.95-1ubuntu6). I check tracker-status, and it reports 'Paused'. But the tracker-applet had 'Pause All Indexing' unmarked. So I selected 'Pause All Indexing', tracker-status still reported 'Paused'. I then unselected 'Pause All Indexing', tracker-status now reports 'Idle', and trackerd goes to sleep.

Next time this happens I will try to collect some more information.

[Raf (4263004-noduck)]

Seems to happen frequently when a lot of new files are created (e.g. extracting a kernel archive).

It is just spinning in g_main_loop_run, with a zero timeout on poll (see strace above). gdb backtrace:
#0 0x00007fb3c6e92e68 in ?? () from /usr/lib/libdbus-glib-1.so.2
#1 0x00007fb3c5e88a81 in g_main_context_prepare () from /lib/libglib-2.0.so.0
#2 0x00007fb3c5e88ee8 in ?? () from /lib/libglib-2.0.so.0
#3 0x00007fb3c5e894bc in g_main_context_iteration ()
   from /lib/libglib-2.0.so.0
#4 0x0000000000426899 in inotify_watch_func (source=<value optimized out>,
    condition=<value optimized out>, data=<value optimized out>)
    at inotify-monitor.c:93
#5 0x00007fb3c5e85492 in g_main_context_dispatch () from /lib/libglib-2.0.so.0
#6 0x00007fb3c5e89308 in ?? () from /lib/libglib-2.0.so.0
#7 0x00007fb3c5e8981d in g_main_loop_run () from /lib/libglib-2.0.so.0
#8 0x000000000041cb16 in main (argc=1, argv=0x7fffc2fc3848)
    at tracker-main.c:1236

[RiaanFV (riaanfventer)]

me 2

[Launchpad Janitor (janitor)]

[Expired for tracker (Ubuntu) because there has been no activity for 60 days.]