Heavy Memory leak in gvfsd-metadata

Bug #459439 reported by Dylan Taylor
26
This bug affects 4 people
Affects Status Importance Assigned to Milestone
gvfs (Ubuntu)
Invalid
Medium
Ubuntu Desktop Bugs

Bug Description

Binary package hint: gvfs

I'm on Ubuntu Karmic 9.10 Beta with all the latest updates. I'm currently having issues with system stability. After messing around with System Monitor, I noticed that when the freezes all occur, the process gvfs-metadata is using around 95% of my Core2 Quad CPU, and the process is quickly increasing in memory, until there is no more left. I have seen gvfs-metadata alone use up 1.8GB of RAM. When it can no longer take up any more ram, the system slows down and freezes up, sometimes automatically terminating the process after 5 minutes, only to randomly start it again later, causing the issue again. Killing the process results in it automatically being restarted, and it once again starts to take up memory. The issue begins to occur shortly after gnome finishes loading, but it does not occur in an xterm session. When the system locks up when this process freezes it, I have noticed that Alt-PrintScr-K does not return me to gdm, but rather, there is corrupt graphics at the top of the screen, and the system needs to be forcefully restarted.

ProblemType: Bug
Architecture: i386
CheckboxSubmission: 060a7c86f0637c4ab3689dae9311a12c
CheckboxSystem: e0ccf4c605a3846e6e356cb4f0643aca
Date: Fri Oct 23 18:11:33 2009
DistroRelease: Ubuntu 9.10
Package: gvfs 1.4.1-0ubuntu1
ProcEnviron:
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.31-14.48-generic
SourcePackage: gvfs
Uname: Linux 2.6.31-14-generic i686
XsessionErrors:
 (gnome-settings-daemon:18941): GLib-CRITICAL **: g_propagate_error: assertion `src != NULL' failed
 (nautilus:19037): Eel-CRITICAL **: eel_preferences_get_boolean: assertion `preferences_is_initialized ()' failed
 (polkit-gnome-authentication-agent-1:19061): GLib-CRITICAL **: g_once_init_leave: assertion `initialization_value != 0' failed
 (gnome-panel:19036): Gdk-WARNING **: /build/buildd/gtk+2.0-2.18.3/gdk/x11/gdkdrawable-x11.c:952 drawable is not a pixmap or window
 (gnome-appearance-properties:19186): Gdk-CRITICAL **: gdk_display_sync: assertion `GDK_IS_DISPLAY (display)' failed

Revision history for this message
Dylan Taylor (dylanmtaylor) wrote :
Revision history for this message
Chris Coulson (chrisccoulson) wrote :

Thank you for taking the time to report this bug and helping to make Ubuntu better. Please try to obtain a valgrind log following the instructions at https://wiki.ubuntu.com/Valgrind and attach the file to the bug report. This will greatly help us in tracking down your problem.

Changed in gvfs (Ubuntu):
status: New → Incomplete
assignee: nobody → Ubuntu Desktop Bugs (desktop-bugs)
importance: Undecided → Medium
Revision history for this message
Dylan Taylor (dylanmtaylor) wrote :

Chris - This bug is very important to me, as it makes my system virtually unusable at times. I will attempt to get you any information you need.

Revision history for this message
Dylan Taylor (dylanmtaylor) wrote :

I just noticed that the name of the process was gvfsd-metadata instead of gvfs-metadata. I have changed the description accordingly. Also, I am having trouble getting that valgrind log, although the issue is definately still occuring. I wish to help as much as I can, so could you help me get you the information you need?

summary: - Heavy Memory leak in gvfs-metadata
+ Heavy Memory leak in gvfsd-metadata
Revision history for this message
Dylan Taylor (dylanmtaylor) wrote :

I have determined that the memory leaks start roughly 35 seconds after the gnome session is started. The bug occurs every time. I have two other systems running Ubuntu, one is a Asus Eee 1000HA netbook on Jaunty, the other is a 3GB Ram, Core i7 machine also running Karmic Beta, and both of those do not have the same issue. Would there be a way to check for differences in the configurations of the machines to see exactly what is causing the issue?

Revision history for this message
Chris Coulson (chrisccoulson) wrote :

This might have the same underlying issue as bug 424043. That would leak a lot of memory certainly before it crashed anyway, with the repeated mmap calls

Changed in gvfs (Ubuntu):
status: Incomplete → Invalid
Revision history for this message
Dylan Taylor (dylanmtaylor) wrote :

A compressed copy of the gvfs-metadata folder in ~/local/share. This might help debug the problem...

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.