GNOME Bugzilla – Bug 628984
gnome-keyring eating 100% cpu
Last modified: 2010-12-24 14:14:43 UTC
I've had several reports in Fedora against 2.31.91 that say gnome-keyring-daemon is using 100% cpu. Here is one: https://bugzilla.redhat.com/show_bug.cgi?id=630988
Here is another one with some details: https://bugzilla.redhat.com/show_bug.cgi?id=630253 syslog excerpts: Sep 4 14:22:28 meryl gnome-keyring-daemon[3707]: couldn't connect to control socket at: /tmp/keyring-3h2BuG/control: Connection refused Sep 4 14:22:33 meryl kernel: DMA-API: debugging out of memory - disabling Sep 4 14:29:07 meryl gnome-keyring-prompt: Gtk: Failed to load module "atk-bridge": libatk-bridge.so: cannot open shared object file: No such file or directory Sep 4 14:29:08 meryl gnome-keyring-prompt: Bonobo: Bonobo must be initialized before use Sep 4 14:29:18 meryl gnome-keyring-daemon[3707]: couldn't allocate secure memory to keep passwords and or keys from being written to the disk Sep 4 14:29:19 meryl gnome-keyring-prompt: Gtk: Failed to load module "atk-bridge": libatk-bridge.so: cannot open shared object file: No such file or directory Sep 4 14:29:19 meryl gnome-keyring-prompt: Bonobo: Bonobo must be initialized before use
Is there any way, we can get a stack trace so we can identify which part of the daemon this is happening in? For anyone listening in. To get a stack trace of a 100% CPU process you follow the normal procedure for stack traces: http://live.gnome.org/GettingTraces But since the process hasn't crashed, you induce a crash by sending a send the process a ABRT, TRAP, or signal. For example: $ killall -TRAP gnome-keyring-daemon This will normally induce a crash. For a 100% CPU hung process we need several stack traces to pin point the area where it's looping around in.
Is this the same bug as bug #624254
It might be... is there an easy way to duplicate? I've been waiting for the problem to occur again on my system so that I could try and grab a backtrace but nothing so far.
To duplicate bug #624254 a process (like socat) connects to the GPG agent unix socket and then diconnects. However duplicating that bug, doesn't help with determining whether this bug is fixed.
Do we want to keep the bug open, it's been a long time since there is any update or feedback from user ?
Good plan Akhil: Closing this bug report as no further information has been provided. Please feel free to reopen this bug if you can provide the information asked for. Thanks!