Description of problem: Happened right after logging in into X. Version-Release number of selected component: mate-file-manager-1.6.4-1.fc20 Additional info: reporter: libreport-2.2.0 backtrace_rating: 4 cmdline: caja --sm-client-id 105468dfe7edd118b3138869692717713600000016210027 --sm-client-state-file /home/jin/.config/session-state/caja-1396794031.desktop crash_function: _g_log_abort executable: /usr/bin/caja kernel: 3.13.8-200.fc20.x86_64 runlevel: N 5 type: CCpp uid: 1000 Truncated backtrace: Thread no. 1 (10 frames) #2 _g_log_abort at gmessages.c:255 #5 caja_file_new_from_filename at caja-file.c:505 #6 caja_file_get_internal at caja-file.c:696 #7 caja_file_get_by_uri at caja-file.c:743 #8 caja_directory_get_corresponding_file at caja-directory.c:463 #9 begin_location_change at caja-window-manage-views.c:982 #10 caja_window_slot_open_location_full at caja-window-manage-views.c:690 #11 caja_window_go_to at caja-window.c:241 #12 caja_application_load_session at caja-application.c:2294 #13 caja_application_startup at caja-application.c:1067
Created attachment 883371 [details] File: backtrace
Created attachment 883372 [details] File: cgroup
Created attachment 883373 [details] File: core_backtrace
Created attachment 883374 [details] File: dso_list
Created attachment 883375 [details] File: environ
Created attachment 883376 [details] File: limits
Created attachment 883377 [details] File: maps
Created attachment 883378 [details] File: open_fds
Created attachment 883379 [details] File: proc_pid_status
Created attachment 883380 [details] File: var_log_messages
Can you describe a little more detailed what happend? Did this issue occurs frequently ?
Well, unfortunately I can't really say how to reproduce. I am pretty sure the first time it happened was after the Fedora update which installed mate-file-manager 1.6.4-2.fc20.x86_64, I did not have this problem before. I usually hibernate but at some point I restarted the system after a kernel update and right after logging in into X I got the crash notification, I did not do anything at that point yet, was actually just waiting for everything to load. I then logged out and logged in again and got the same result, then while being logged in continued to crash randomly, although not very often. Later I figured that I had another problem, caja seemed to be hogging CPU at 99%, killing it helped for only a short time. So then I unchecked the "Start Caja desktop at log in" in the mate-session-properties and the problem disappeared. From that point on I also did not have any more crashes, i.e. the problem described in the ticket was gone after the above change. Caja is running though, I do see it in the process list. Hope that helps, otherwise if you have anything special that you would like me to try, please let me know.
(In reply to Sergey Bostandzhyan from comment #12) > Well, unfortunately I can't really say how to reproduce. I am pretty sure > the first time it happened was after the Fedora update which installed > mate-file-manager 1.6.4-2.fc20.x86_64, I did not have this problem before. I > usually hibernate but at some point I restarted the system after a kernel > update and right after logging in into X I got the crash notification, I did > not do anything at that point yet, was actually just waiting for everything > to load. > > I then logged out and logged in again and got the same result, then while > being logged in continued to crash randomly, although not very often. > > Later I figured that I had another problem, caja seemed to be hogging CPU at > 99%, killing it helped for only a short time. So then I unchecked the "Start > Caja desktop at log in" in the mate-session-properties and the problem > disappeared. From that point on I also did not have any more crashes, i.e. > the problem described in the ticket was gone after the above change. > > Caja is running though, I do see it in the process list. > > Hope that helps, otherwise if you have anything special that you would like > me to try, please let me know. Ok, this explains a lot. With mate-file-manager-1.6.4-2.fc20.x86_64 and mate-desktop-1.6.2-2.fc20 i switched back to mate defaults to start caja with mate-session, and not using the caja-autostart script anymore. The caja-autostart script was needed to workaround the horrible x-caja-windows issue in the past with a start delay. But since upstream fix the issue using a delay isn't needed anymore. Note, both updates need to be apply together, because mate-desktop-1.6.2-2.fc20 sets the gsettings settings for starting caja back to default. I quess both updates wouldn't applied together and this caused the issue. Sorry, i did test updatingthose packages on severalmachines and VM's, here all was running fine. Also both updates are in the same push to repos. https://admin.fedoraproject.org/updates/FEDORA-2014-4710/mate-desktop-1.6.2-2.fc20,mate-file-manager-1.6.4-2.fc20?_csrf_token=12dcb19cf322d4259ef57bc76ef9a0239a683a6c OK, you can delete ~.config/autostart/caja-autostart.desktop. And you should have this gsettings output in a terminal [rave@mother ~]$ gsettings get org.mate.session.required-components filemanager 'caja' I hope you agree to close this issue ;)
I always update via "yum update" so I guess whatever was in the update got installed together. I followed your instructions and re-logged in afterwards - everything seems to be OK now, so yes, you can close it :) Thanks for looking into it.
Unfortunately I have problems again, interestingly enough abrt trace points me to this very bug which we have closed. Bboot fresh, log into X, do nothing yet - and I get a mate file manager crash reported by abrtd. get org.mate.session.required-components filemanager still returns 'caja' and I also do not have the caja-autostart.desktop file which we deleted last time. So really no clue where the problem is coming from... Please let me know if I can debug something for you.
That means caja want's to start but it crashed and you have no right click function for the desktop? There should be a new generated backtrace under /var/tmp/abrt/<issue-folder> Better you delete all old folders there and let abrt create a new one there. And attach it here.
....and now it's not happening anymore, although I could reproduce it consistently before. There were some updates in between but I think only for non-mate components. Not sure if this had any impact. I guess we can close this one again? :P