It seems to crash when I switch windows (alt+tab ; even falling back to the static application switcher does not alleviate the bug).
For reference, I do a lot of translation work, which requires me to have gedit, mplayer, evince running at the same time and heavily switching between these. (Especially mplayer/evince probably take a lot of memory). When metacity crashes, everything freezes while apport runs and writes its files, during which only the audio players and non graphical stuff keep running (if I am lucky I can SSH and force gedit to save my work), and once apport finishes, in a split second, every window just loses its decorations, and I am forced back to the login screen. Logs indeed indicate this : /var/log/syslog.1:Oct 20 05:22:14 archer gnome-session[29421]: WARNING: App 'metacity.desktop' respawning too quickly /var/log/syslog.1:Oct 20 05:22:14 archer gnome-session[29421]: gnome-session[29421]: WARNING: App 'metacity.desktop' respawning too quickly But I can't confirm with the bare eye that it tries to reload anything, or what the respawn threshold is. -- You received this bug notification because you are a member of Desktop Packages, which is subscribed to metacity in Ubuntu. https://bugs.launchpad.net/bugs/1507764 Title: metacity crashed with signal 5 in _XEventsQueued() ; forcing gnome session to logout Status in metacity package in Ubuntu: New Bug description: This bug seemed to occur already since earlier versions of Ubuntu, and usually would just end up leaving me with undecorated windows, from which recovery was as easy as restarting metacity by hand, but in this release of Ubuntu (15.04), it seems gnome-session tries to restart it too many times in a too short span, and then decides arbitrarily that it failed so bad it has to kill the whole session. So, while the metacity bug is really a thing, my real problem actually might be a gnome-session-bin bug. I expect metacity crashing would not bring down the whole session, as is the case currently. At any rate, this is extremely annoying, I am hitting this bug about every two days, and because of it and the regression in crash handling, I end up losing all my windows and running programs. (In all fairness, I am suspecting a hardware issue, but the problem remains that a sub process dying should not cause the parent process to die like this) ProblemType: Crash DistroRelease: Ubuntu 15.04 Package: metacity 1:3.14.3-1ubuntu7 ProcVersionSignature: Ubuntu 3.19.0-31.36-generic 3.19.8-ckt7 Uname: Linux 3.19.0-31-generic x86_64 NonfreeKernelModules: nvidia ApportVersion: 2.17.2-0ubuntu1.5 Architecture: amd64 CrashCounter: 1 CurrentDesktop: GNOME-Flashback:Unity Date: Tue Oct 20 05:19:37 2015 ExecutablePath: /usr/bin/metacity InstallationDate: Installed on 2010-01-30 (2088 days ago) InstallationMedia: Ubuntu 9.10 "Karmic Koala" - Release amd64 (20091027.1) ProcCmdline: metacity ProcEnviron: LANGUAGE=en_US:en PATH=(custom, user) XDG_RUNTIME_DIR=<set> LANG=en_US.UTF-8 SHELL=/bin/zsh Signal: 5 SourcePackage: metacity StacktraceTop: ?? () from /usr/lib/x86_64-linux-gnu/libX11.so.6 ?? () from /usr/lib/x86_64-linux-gnu/libX11.so.6 _XEventsQueued () from /usr/lib/x86_64-linux-gnu/libX11.so.6 _XGetRequest () from /usr/lib/x86_64-linux-gnu/libX11.so.6 XFixesCopyRegion () from /usr/lib/x86_64-linux-gnu/libXfixes.so.3 Title: metacity crashed with signal 5 in _XEventsQueued() UpgradeStatus: Upgraded to vivid on 2015-09-05 (44 days ago) UserGroups: adm admin audio cdrom dialout libvirtd lpadmin plugdev pulse pulse-access sambashare vboxusers video To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/metacity/+bug/1507764/+subscriptions -- Mailing list: https://launchpad.net/~desktop-packages Post to : [email protected] Unsubscribe : https://launchpad.net/~desktop-packages More help : https://help.launchpad.net/ListHelp
