Summary: | kglobalaccel5 repeatedly crashes on boot in QGuiApplicationPrivate::createPlatformIntegration(), which can delay login | ||
---|---|---|---|
Product: | [Frameworks and Libraries] frameworks-kglobalaccel | Reporter: | mintjulep <guillaume.gayot> |
Component: | general | Assignee: | kdelibs bugs <kdelibs-bugs> |
Status: | RESOLVED FIXED | ||
Severity: | major | CC: | cirlo_ca, fabian, flateric, i, kde, kde, med.medin.2014, mr-peipei, nate, plasma-bugs, windows2linux |
Priority: | NOR | ||
Version: | 5.80.0 | ||
Target Milestone: | --- | ||
Platform: | Arch Linux | ||
OS: | Linux | ||
URL: | https://bugzilla.redhat.com/show_bug.cgi?id=1787230 | ||
See Also: | https://bugs.kde.org/show_bug.cgi?id=450320 | ||
Latest Commit: | https://commits.kde.org/kglobalaccel/25e84f3f6ac54097d9c021705dc18b187352fa02 | Version Fixed In: | Plasma 5.22 or Frameworks 5.82 |
Description
mintjulep
2020-11-21 01:39:49 UTC
JourLancel show errors related to kglobalaccel. Hi, systemctl --user mask plasma-kglobalaccel.service Hi, Hi, please allow me to amend on this bug report. The problem is that an immediate login after logout will cause stuck in Plasma. From the logs, we see that it's the failure to start kglobalaccel that causes the problem. After the timeout, Plasma starts and kglobalaccel also starts successfully. Related systemd logs: dbus-daemon[1298]: [session uid=1000 pid=1298] Activating via systemd: service name='org.kde.kglobalaccel' unit='plasma-kglobalaccel.service' requested by ':1.597' (uid=1000 pid=44701 comm="/usr/bin/ksmserver ") systemd[1067]: plasma-kglobalaccel.service: Start request repeated too quickly. systemd[1067]: plasma-kglobalaccel.service: Failed with result 'core-dump'. ksmserver[44701]: Couldn't start kglobalaccel from org.kde.kglobalaccel.service: QDBusError("org.freedesktop.DBus.Error.NoReply", "Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.") kded5[44696]: Couldn't start kglobalaccel from org.kde.kglobalaccel.service: QDBusError("org.freedesktop.DBus.Error.NoReply", "Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.") kaccess[44719]: Couldn't start kglobalaccel from org.kde.kglobalaccel.service: QDBusError("org.freedesktop.DBus.Error.NoReply", "Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.") org_kde_powerdevil[44816]: Couldn't start kglobalaccel from org.kde.kglobalaccel.service: QDBusError("org.freedesktop.DBus.Error.NoReply", "Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.") dbus-daemon[1298]: [session uid=1000 pid=1298] Failed to activate service 'org.kde.kglobalaccel': timed out (service_start_timeout=120000ms) kactivitymanagerd[44827]: Couldn't start kglobalaccel from org.kde.kglobalaccel.service: QDBusError("org.freedesktop.DBus.Error.TimedOut", "Failed to activate service 'org.kde.kglobalaccel': timed out (service_start_timeout=120000ms)") dbus-daemon[1298]: [session uid=1000 pid=1298] Activating via systemd: service name='org.kde.kglobalaccel' unit='plasma-kglobalaccel.service' requested by ':1.610' (uid=1000 pid=44816 comm="/usr/lib/org_kde_powerdevil ") dbus-daemon[1298]: [session uid=1000 pid=1298] Successfully activated service 'org.kde.kglobalaccel' kglobalaccel5[44876]: The X11 connection broke (error 1). Did the X11 server die? systemd[1067]: plasma-kglobalaccel.service: Main process exited, code=exited, status=1/FAILURE systemd[1067]: plasma-kglobalaccel.service: Failed with result 'exit-code'. We have seen similar reports in multi distros: - https://bugzilla.redhat.com/show_bug.cgi?id=1787230 - https://forums.gentoo.org/viewtopic-t-1046222-start-0.html - https://forum.manjaro.org/t/cant-login-after-logout/40887 Sorry for the repeated comment submission. (In reply to outloudvi from comment #2) > systemctl --user mask plasma-kglobalaccel.service It seems that masking the kglobalaccel service will solve the stuck problem. After the plasmashell started, I ran kglobalaccel5 manually and global key shortcuts work as usual. I think that there are two reasons for the problem, but not sure which: 1. kglobalaccel5 should not be started before plasmashell, but dbus asks it to do so, which results in the problem. 2. kglobalaccel5 is meant to be started before plasmashell, but it does not work now because of some problems. commit 25e84f3f6ac54097d9c021705dc18b187352fa02 Author: David Edmundson <kde@davidedmundson.co.uk> Date: Wed Dec 2 14:57:27 2020 +0000 Avoid autostarting kglobalaccel when shutting down in kglobalaccel should hide the problem triggering this. That'll be in the next frameworks 5.77. How would I do that ? You would upgrade to Frameworks 5.77 when your distro makes it available. Hum... after checking, I'm allready on kglobalaccel 5.77 :/ I'm on Manjaro with frameworks 5.77.0 and get the same problem with kglobalaccel. David, does this require changes in Plasma 5.21 too? I am suffering from the same issue as described by the bug report. If I logout and then try to log back in, the SDDM screen will freeze for about three minutes before finally logging in to a new Plasma session. Disabling Compositing made no difference. Using TripleBuffering made no difference. Removing ~/.cache/ seems to help sometimes, but other times it makes no difference. Not very consistent. Removing ~/.Xauthority made no difference. Here is my system info, and let me know what logs/debug I can provide to help diagnose and fix this bug: Operating System: Manjaro Linux KDE Plasma Version: 5.21.3 KDE Frameworks Version: 5.80.0 Qt Version: 5.15.2 Kernel Version: 5.10.23-1-MANJARO OS Type: 64-bit Graphics Platform: X11 Processors: 6 × Intel® Core™ i5-8400 CPU @ 2.80GHz Memory: 11.6 GiB of RAM Graphics Processor: GeForce GTX 1650/PCIe/SSE2 Nvidia Driver Version: 460.56 (proprietary) I cannot recall how long this has been a problem, but I tried to ignore it in hopes that maybe it's already been addressed and will be fixed with an update to KDE. I would say I've noticed it for months, perhaps. > systemctl --user mask plasma-kglobalaccel.service
If I do this, then it seems to fix the login delay problem, but now I cannot use *any* keyboard shortcuts. :(
(In reply to flan_suse from comment #13) > > systemctl --user mask plasma-kglobalaccel.service > > If I do this, then it seems to fix the login delay problem, but now I cannot > use *any* keyboard shortcuts. :( That's expected as kglobalaccel is required for global keyboard shortcuts. Also, I want to amend that this bug is still there at the follow version: * Plasma 5.21.3 * KDE Framework 5.80.0 * kglobalaccel 5.80.0 (Arch Linux 5.80.0:2, [1]) Also I noticed that kglobalaccel crashed before that failure of KDE re-login: Process 4312 (kglobalaccel5) of user 1000 dumped core. Stack trace of thread 4312: #0 0x00007faaea8c4ef5 raise (libc.so.6 + 0x3cef5) #1 0x00007faaea8ae862 abort (libc.so.6 + 0x26862) #2 0x00007faaeacc49ac _ZNK14QMessageLogger5fatalEPKcz (libQt5Core.so.5 + 0x909ac) #3 0x00007faaeb2ba646 _ZN22QGuiApplicationPrivate25createPlatformIntegrationEv (libQt5Gui.so.5 + 0x131646) #4 0x00007faaeb2baae1 _ZN22QGuiApplicationPrivate21createEventDispatcherEv (libQt5Gui.so.5 + 0x131ae1) #5 0x00007faaeaef0686 _ZN23QCoreApplicationPrivate4initEv (libQt5Core.so.5 + 0x2bc686) #6 0x00007faaeb2bdb20 _ZN22QGuiApplicationPrivate4initEv (libQt5Gui.so.5 + 0x134b20) #7 0x00007faaeb2bea88 _ZN15QGuiApplicationC1ERiPPci (libQt5Gui.so.5 + 0x135a88) #8 0x000056524a34b08f n/a (kglobalaccel5 + 0x208f) #9 0x00007faaea8afb25 __libc_start_main (libc.so.6 + 0x27b25) #10 0x000056524a34b4de n/a (kglobalaccel5 + 0x24de) 1. https://github.com/archlinux/svntogit-packages/blob/c1075e2b63a192d98d8ef19ab6a242edd14e6fce/trunk/PKGBUILD Can we safely change this bug's status to "Confirmed" and remove that it has been "Fixed" for KDE Framework 5.77, since it still exists in 5.80? Are you using systemd startup, or not? See also http://blog.davidedmundson.co.uk/blog/plasma-and-the-systemd-startup/ (In reply to Nate Graham from comment #16) > Are you using systemd startup, or not? See also > http://blog.davidedmundson.co.uk/blog/plasma-and-the-systemd-startup/ sweet jesus lord of mercy it worked !!! what I did : kwriteconfig5 --file startkderc --group General --key systemdBoot true then check the status with : systemctl --user status plasma-plasmashell‧service and saw that the service was inactive. I enabled it : systemctl --user enable plasma-plasmashell‧service reboot and it's perfect ! Not only can I now logout and login without any problem, but also Plasma starts faster than before Thanks for the hint !! EDIT : No it doesn't, talk to fast. the bug came back. Not the first time I get this tho : it's hard to know what's up because sometimes, it doesn't bug Yay! ... almost lol. Can you see if it's still kglobalaccel that's crashing on boot, or is it something else now? `coredumpctl` should be able to tell you. Here is the result of cordumpctl : Thu 2020-10-22 19:18:43 CEST 1026 1000 985 SIGABRT missing /usr/bin/plasmashell n/a Thu 2020-10-22 19:18:43 CEST 1027 1000 985 SIGABRT missing /usr/lib/drkonqi n/a Thu 2020-10-22 19:18:43 CEST 999 1000 985 SIGABRT missing /usr/bin/kglobalaccel5 n/a Thu 2020-10-22 19:18:46 CEST 713 1000 985 SIGSEGV missing /usr/bin/plasmashell n/a Thu 2020-10-22 19:26:00 CEST 2659 1000 985 SIGABRT missing /usr/bin/plasmashell n/a Thu 2020-10-22 19:26:00 CEST 2660 1000 985 SIGABRT missing /usr/lib/drkonqi n/a Thu 2020-10-22 19:26:00 CEST 2629 1000 985 SIGABRT missing /usr/bin/kglobalaccel5 n/a Thu 2020-10-22 19:26:03 CEST 1206 1000 985 SIGSEGV missing /usr/bin/plasmashell n/a Thu 2020-10-22 19:30:03 CEST 4149 1000 985 SIGABRT missing /usr/bin/kglobalaccel5 n/a Thu 2020-10-22 19:30:03 CEST 4184 1000 985 SIGABRT missing /usr/lib/drkonqi n/a Thu 2020-10-22 19:30:03 CEST 4183 1000 985 SIGABRT missing /usr/bin/plasmashell n/a Thu 2020-10-22 19:30:06 CEST 2837 1000 985 SIGSEGV missing /usr/bin/plasmashell n/a Thu 2020-10-22 19:43:44 CEST 662 1000 985 SIGSEGV missing /usr/bin/kded5 n/a Thu 2020-10-22 20:31:49 CEST 26836 1000 985 SIGABRT missing /usr/lib/drkonqi n/a Thu 2020-10-22 20:31:49 CEST 26835 1000 985 SIGABRT missing /usr/bin/plasmashell n/a Thu 2020-10-22 20:31:53 CEST 708 1000 985 SIGSEGV missing /usr/bin/plasmashell n/a Thu 2020-10-22 21:00:36 CEST 17077 1000 985 SIGABRT missing /usr/bin/plasmashell n/a Thu 2020-10-22 21:00:36 CEST 17078 1000 985 SIGABRT missing /usr/lib/drkonqi n/a Thu 2020-10-22 21:00:39 CEST 715 1000 985 SIGSEGV missing /usr/bin/plasmashell n/a Thu 2020-10-22 21:12:39 CEST 988 1000 985 SIGSEGV missing /usr/bin/kdeinit5 n/a Thu 2020-10-22 21:29:39 CEST 10467 1000 985 SIGABRT missing /usr/lib/drkonqi n/a Thu 2020-10-22 21:29:39 CEST 10466 1000 985 SIGABRT missing /usr/bin/plasmashell n/a Thu 2020-10-22 21:29:42 CEST 717 1000 985 SIGSEGV missing /usr/bin/plasmashell n/a Thu 2020-10-22 21:40:50 CEST 12881 1000 985 SIGABRT missing /usr/lib/drkonqi n/a wich is weird cause we're not Tuesday 2020/10/22 Ok, made coredumpctl -r and got this : Mon 2021-04-05 20:36:22 CEST 2753 1000 985 SIGABRT present /usr/bin/kglobalaccel5 403.1K Mon 2021-04-05 20:36:21 CEST 2739 1000 985 SIGABRT present /usr/bin/kglobalaccel5 403.4K Mon 2021-04-05 20:36:21 CEST 2701 1000 985 SIGABRT present /usr/bin/kglobalaccel5 403.2K Mon 2021-04-05 20:36:20 CEST 2681 1000 985 SIGABRT present /usr/bin/kglobalaccel5 417.0K Mon 2021-04-05 20:36:20 CEST 2665 1000 985 SIGABRT present /usr/bin/kglobalaccel5 410.3K Mon 2021-04-05 20:32:15 CEST 2172 1000 985 SIGABRT present /usr/bin/kglobalaccel5 402.9K Mon 2021-04-05 20:32:15 CEST 2158 1000 985 SIGABRT present /usr/bin/kglobalaccel5 403.0K Mon 2021-04-05 20:32:14 CEST 2120 1000 985 SIGABRT present /usr/bin/kglobalaccel5 402.9K Mon 2021-04-05 20:32:13 CEST 2099 1000 985 SIGABRT present /usr/bin/kglobalaccel5 403.2K Mon 2021-04-05 20:29:17 CEST 2427 1000 985 SIGABRT present /usr/bin/kglobalaccel5 403.6K Mon 2021-04-05 20:29:17 CEST 2413 1000 985 SIGABRT present /usr/bin/kglobalaccel5 403.5K Mon 2021-04-05 20:29:16 CEST 2375 1000 985 SIGABRT present /usr/bin/kglobalaccel5 402.9K Mon 2021-04-05 20:29:15 CEST 2354 1000 985 SIGABRT present /usr/bin/kglobalaccel5 403.2K so yeah, seems like the same issue Darn. Can you attach a backtrace from one of those new crashlogs? `coredumpctl debug kglobalaccel5` and then type `bt` [Thread debugging using libthread_db enabled] Using host libthread_db library "/usr/lib/libthread_db‧so.1". Core was generated by `/usr/bin/kglobalaccel5'. Program terminated with signal SIGABRT, Aborted. #0 0x00007f1052a57ef5 in raise () from /usr/lib/libc‧so.6 (gdb) bt #0 0x00007f1052a57ef5 in raise () from /usr/lib/libc‧so.6 #1 0x00007f1052a41862 in abort () from /usr/lib/libc‧so.6 #2 0x00007f1052e579ac in QMessageLogger::fatal(char const*, ...) const () from /usr/lib/libQt5Core‧so.5 #3 0x00007f105344d646 in QGuiApplicationPrivate::createPlatformIntegration() () from /usr/lib/libQt5Gui‧so.5 #4 0x00007f105344dae1 in QGuiApplicationPrivate::createEventDispatcher() () from /usr/lib/libQt5Gui‧so.5 #5 0x00007f1053083686 in QCoreApplicationPrivate::init() () from /usr/lib/libQt5Core‧so.5 #6 0x00007f1053450b20 in QGuiApplicationPrivate::init() () from /usr/lib/libQt5Gui‧so.5 #7 0x00007f1053451a88 in QGuiApplication::QGuiApplication(int&, char**, int) () from /usr/lib/libQt5Gui‧so.5 #8 0x000055d528b2308f in ?? () #9 0x00007f1052a42b25 in __libc_start_main () from /usr/lib/libc‧so.6 #10 0x000055d528b234de in ?? () (gdb) coredumpctl -r yields the same output as @mintjulop above, even after applying the kwriteconfig / systemdboot workaround. After multiple reboots (to be certain), the re-login issue remains, even after confirming that "plasma-plasmashell.service" is enabled and active. I tried to follow the steps to provide a backtrace, but I get this warning: "Coredump entry has no core attached (neither internally in the journal nor externally on disk)." My system info once again is: Operating System: Manjaro Linux KDE Plasma Version: 5.21.3 KDE Frameworks Version: 5.80.0 Qt Version: 5.15.2 Kernel Version: 5.10.23-1-MANJARO OS Type: 64-bit Graphics Platform: X11 Processors: 6 × Intel® Core™ i5-8400 CPU @ 2.80GHz Memory: 11.6 GiB of RAM Graphics Processor: GeForce GTX 1650/PCIe/SSE2 Nvidia Driver Version: 460.56 (proprietary) I think if this resolves for @mintjulep, it will resolve for others, including myself. I don't want to distract nor interrupt anything, but if providing another coredump or backtrace would be useful, let me know. So far, what @mintjulep is pasting looks nearly identical to my own terminal output. If it means anything, this is what I get from "coredumpctl debug kglobalaccel5": Stack trace of thread 8649: #0 0x00007fe650683ef5 raise (libc.so.6 + 0x3cef5) #1 0x00007fe65066d862 abort (libc.so.6 + 0x26862) #2 0x00007fe650a819ac _ZNK14QMessageLogger5fatalEPKcz (libQt5Core.so.5 + 0x909ac) #3 0x00007fe651079646 _ZN22QGuiApplicationPrivate25createPlatformIntegrationEv (libQt5Gui.so.5 + 0x131646) #4 0x00007fe651079ae1 _ZN22QGuiApplicationPrivate21createEventDispatcherEv (libQt5Gui.so.5 + 0x131ae1) #5 0x00007fe650cad686 _ZN23QCoreApplicationPrivate4initEv (libQt5Core.so.5 + 0x2bc686) #6 0x00007fe65107cb20 _ZN22QGuiApplicationPrivate4initEv (libQt5Gui.so.5 + 0x134b20) #7 0x00007fe65107da88 _ZN15QGuiApplicationC1ERiPPci (libQt5Gui.so.5 + 0x135a88) #8 0x000055b77a53408f n/a (kglobalaccel5 + 0x208f) #9 0x00007fe65066eb25 __libc_start_main (libc.so.6 + 0x27b25) #10 0x000055b77a5343fe n/a (kglobalaccel5 + 0x23fe) Thanks everyone. >but if providing another coredump or backtrace would be useful, let me know
Not really. We know why kglobalaccel fails if it gets started without X running.
What would be interesting to know is why kglobalaccel is getting started/restarted at this point.
What I believe is happening is kglobalaccel has X11 torn away from under it (which is not too surprising) and fails with an error code, it then gets restarted by systemd because a unit exists. We don't have an X server so crash 5 times in a loop - we then get blacklisted and it fails to start when X11 really is up. With the full systemd session the PartOf=GraphicalSession.target somewhat helps I can reproduce this happening, but can't immediately manage to reproduce any ill-effect, it will be timing related. I've thought of two remedies that I think will both work. I would like it if someone is in a position to verify. Solution 1: Edit plasma-kglobalaccel.service and add to the service: SuccessExitStatus=0 1 This will mean on the first time that the X11 connection is pulled out from under it, we won't try and restart. Solution 2: During plasma boot, call: systemctl --user reset-failed early in the boot, like in /etc/X11/Xsession.d I would like both verified and we'll merge an implementation of both. *** Bug 435486 has been marked as a duplicate of this bug. *** (In reply to David Edmundson from comment #29) > I would like it if someone is in a position to verify. > > Solution 1: > Edit plasma-kglobalaccel.service and add to the service: > SuccessExitStatus=0 1 > > > This will mean on the first time that the X11 connection is pulled out from > under it, we won't try and restart. This appears to work. I tried logging in and out and in and out and in and out as quickly as possible, yet it never hangs or freezes (let alone just sit at the SDDM greeter login screen for 3 minutes). coredumpctl -r shows no entries for kglobalaccel5 either. (Some new ones for "org_kde_powerdevil" appeared, but that might just be a coincidence. (I hope!) I'll try out Solution #2 as well. (In reply to David Edmundson from comment #29) > Solution 2: > > During plasma boot, call: > systemctl --user reset-failed > > early in the boot, like in /etc/X11/Xsession.d > > > I would like both verified and we'll merge an implementation of both. I undid the changes from Solution #1, and then followed Solution #2 as closely as possible. 1) I made an executable file named /etc/X11/xinit/xinitrc.d/99-reset-failed-user-service.sh 2) I filled it with the following: #!/bin/sh systemctl --user reset-failed This also appears to solve the issue. The only difference is even though there is no apparent delay or freeze when re-logging in over and over and over, Solution #2 still populates "coredumpctl -r" with entries for "kglobalaccel5". Not sure if that means anything. So in summary, Solution #1 and Solution #2 both seem to work, no matter how often nor how quickly I re-login. :) The only caveat is there are still core dumps for kglobalaccel5 when using Solution #2. Ok, as for me, the first one didn't work : login logout make the same issue. It finaly start, after a very long time. The second option is kind of out of reach for me, as I don't know how to call systemctl --user reset-failed during boot Thanks, I'll turn those into merge requests on all our services. I am going for the double-fix so we handle it super safely https://invent.kde.org/frameworks/kglobalaccel/-/merge_requests/17 https://invent.kde.org/plasma/plasma-workspace/-/merge_requests/773 Sweet ! When do you think it would be updated ? (In reply to mintjulep from comment #33) > Ok, as for me, the first one didn't work : login logout make the same issue. > It finaly start, after a very long time. > > The second option is kind of out of reach for me, as I don't know how to call > systemctl --user reset-failed during boot Did you try out the method I used for Solution #2? Since you're running Arch and I'm on Manjaro, we should share similar base systems. :) As for Solution #1, which file did you edit? (In reply to David Edmundson from comment #35) > I am going for the double-fix so we handle it super safely > > https://invent.kde.org/frameworks/kglobalaccel/-/merge_requests/17 > > https://invent.kde.org/plasma/plasma-workspace/-/merge_requests/773 We are truly grateful to you devs! This was a near show-stopper bug for a common task (re-logging back into Plasma.) Since I'm no programmer nor engineer, and my little brain barely grasps this stuff, I want to ask: are the KDE developers still going to investigate why kglobalaccel5 has been behaving this way for Plasma 5? > What would be interesting to know is why kglobalaccel is getting started/restarted at this point. Even with the above two fixes being applied, they seem to only address errant systemd units (i.e, plasma-kglobalaccel5.service), yet is the real issue deeper with kglobalaccel5 executable / service itself? Correct me if I'm wrong, but in the meantime with these fixes, kglobalaccel5 is still behaving wonky at logout when X11 "is pulled out from under it" like a magic trick with a table cloth. Is it possible this can carry over to Plasma 6 and rear it's ugly head again? (In reply to flan_suse from comment #37) > (In reply to mintjulep from comment #33) > > Ok, as for me, the first one didn't work : login logout make the same issue. > > It finaly start, after a very long time. > > > > The second option is kind of out of reach for me, as I don't know how to call > > systemctl --user reset-failed during boot > > Did you try out the method I used for Solution #2? Since you're running Arch > and I'm on Manjaro, we should share similar base systems. :) > > As for Solution #1, which file did you edit? I'm not a very experienced user so here's what I did : systemctl edit --user plasma-kglobalaccel‧service it opened this in vim : ### Editing /home/mint/.config/systemd/user/plasma-kglobalaccel‧service‧d/override‧conf ### Anything between here and the comment below will become the new contents of the file ### Lines below this comment will be discarded ### /usr/lib/systemd/user/plasma-kglobalaccel‧service # [Unit] # Description=KDE Global Shortcuts Server # PartOf=graphical-session‧target # # [Service] # ExecStart=/usr/bin/kglobalaccel5 # BusName=org‧kde‧kglobalaccel # Slice=background‧slice # TimeoutSec=5sec # Restart=on-failure where I added "SuccesExitStatus=0 1 " this way : ### Editing /home/mint/.config/systemd/user/plasma-kglobalaccel‧service‧d/override‧conf ### Anything between here and the comment below will become the new contents of the file SuccesExitStatus=0 1 ### Lines below this comment will be discarded ### /usr/lib/systemd/user/plasma-kglobalaccel‧service # [Unit] # Description=KDE Global Shortcuts Server # PartOf=graphical-session‧target # # [Service] # ExecStart=/usr/bin/kglobalaccel5 # BusName=org‧kde‧kglobalaccel # Slice=background‧slice # TimeoutSec=5sec # Restart=on-failure after checking, /home/mint/.config/systemd/user/plasma-kglobalaccel‧service‧d/override‧conf content is just : SuccesExitStatus=0 1 When I log out log in, it bugged as ever (maybe took a little less time to finally start) The template you are presented with defaults to everything being "commented out" and thus ignored. The file in its entirety should appear as such: ### Editing /home/mint/.config/systemd/user/plasma-kglobalaccel‧service‧d/override‧conf ### Anything between here and the comment below will become the new contents of the file [Unit] Description=KDE Global Shortcuts Server PartOf=graphical-session.target [Service] ExecStart=/usr/bin/kglobalaccel5 BusName=org.kde.kglobalaccel Slice=background.slice TimeoutSec=5sec Restart=on-failure SuccessExitStatus=0 1 ### Lines below this comment will be discarded ### /usr/lib/systemd/user/plasma-kglobalaccel‧service # [Unit] # Description=KDE Global Shortcuts Server # PartOf=graphical-session‧target # # [Service] # ExecStart=/usr/bin/kglobalaccel5 # BusName=org‧kde‧kglobalaccel # Slice=background‧slice # TimeoutSec=5sec # Restart=on-failure (In reply to flan_suse from comment #40) > The template you are presented with defaults to everything being "commented > out" and thus ignored. > > The file in its entirety should appear as such: > > ### Editing > /home/mint/.config/systemd/user/plasma-kglobalaccel‧service‧d/override‧conf > ### Anything between here and the comment below will become the new contents > of the file > > [Unit] > Description=KDE Global Shortcuts Server > PartOf=graphical-session‧target > > [Service] > ExecStart=/usr/bin/kglobalaccel5 > BusName=org‧kde‧kglobalaccel > Slice=background‧slice > TimeoutSec=5sec > Restart=on-failure > SuccessExitStatus=0 1 > > ### Lines below this comment will be discarded > > ### /usr/lib/systemd/user/plasma-kglobalaccel‧service > # [Unit] > # Description=KDE Global Shortcuts Server > # PartOf=graphical-session‧target > # > # [Service] > # ExecStart=/usr/bin/kglobalaccel5 > # BusName=org‧kde‧kglobalaccel > # Slice=background‧slice > # TimeoutSec=5sec > # Restart=on-failure Tried that, kglobalaccell never started and global shortcut where down after reboot :/ Maybe I did something worng ? As for the 2nd solution, I don't know how to create an executable file like you did (and for the record, I don't understand what this solution is about. The first one I get it : we feed a parameter meaning that we don't care either kglobalaccel started or not) if it's any help, here's the log from journal -f at reloging : Apr 08 16:00:31 Indiana systemd[1001]: Failed to start KDE Plasma Workspace. Apr 08 16:00:31 Indiana systemd[1001]: plasma-plasmashell‧service: Failed with result 'timeout'. Apr 08 16:00:31 Indiana systemd[1001]: plasma-plasmashell‧service: start operation timed out. Terminating. Apr 08 16:00:17 Indiana kded5[4063]: 3 Apr 08 16:00:17 Indiana kded5[4063]: 4 Apr 08 16:00:17 Indiana kded5[4063]: Apr 08 16:00:17 Indiana kded5[4063]: kcm_touchpad: Using X11 backend Apr 08 16:00:17 Indiana NetworkManager[519]: <info> [1617890417.1789] agent-manager: agent[f3c3c8c03ee7bb31,:1.223/org‧kde‧plasma‧networkmanagement/1000]: agent registered Apr 08 16:00:17 Indiana kded5[4063]: org‧kde‧bolt‧kded: Couldn't connect to Bolt DBus daemon Apr 08 16:00:17 Indiana kded5[4063]: org‧kde‧libkbolt: Failed to connect to Bolt manager DBus interface: Apr 08 15:59:51 Indiana systemd[1001]: Starting KDE Plasma Workspace... Apr 08 15:59:51 Indiana systemd[1001]: Started KDE Session Management Server. Apr 08 15:59:27 Indiana systemd-timesyncd[506]: Timed out waiting for reply from 217.91.44.17:123 (3.arch‧pool‧ntp.org). Apr 08 15:59:27 Indiana kactivitymanagerd[4093]: Couldn't start kglobalaccel from org‧kde‧kglobalaccel‧service: QDBusError("org.freedesktop.DBus.Error.NoReply", "Did not receive a reply. Possible causes include> Apr 08 15:59:27 Indiana kwin_x11[4065]: Couldn't start kglobalaccel from org‧kde‧kglobalaccel‧service: QDBusError("org.freedesktop.DBus.Error.NoReply", "Did not receive a reply. Possible causes include: the rem> Apr 08 15:59:27 Indiana kded5[4063]: Couldn't start kglobalaccel from org‧kde‧kglobalaccel‧service: QDBusError("org.freedesktop.DBus.Error.NoReply", "Did not receive a reply. Possible causes include: the remote> Apr 08 15:59:26 Indiana ksmserver[4064]: Couldn't start kglobalaccel from org‧kde‧kglobalaccel‧service: QDBusError("org.freedesktop.DBus.Error.NoReply", "Did not receive a reply. Possible causes include: the re> Apr 08 15:59:19 Indiana dhcpcd[527]: wlp5s0: no IPv6 Routers available Apr 08 15:59:18 Indiana kernel: audit: type=1110 audit(1617890358.836:386): pid=4154 uid=0 auid=1000 ses=12 msg='op=PAM:setcred grantors=pam_securetty,pam_shells,pam_faillock,pam_permit,pam_faillock acct="mint"> Apr 08 15:59:18 Indiana kernel: audit: type=1105 audit(1617890358.836:385): pid=4154 uid=0 auid=1000 ses=12 msg='op=PAM:session_open grantors=pam_loginuid,pam_keyinit,pam_limits,pam_unix,pam_permit,pam_mail,pam> Apr 08 15:59:18 Indiana login[4154]: LOGIN ON tty2 BY mint Apr 08 15:59:18 Indiana audit[4154]: CRED_REFR pid=4154 uid=0 auid=1000 ses=12 msg='op=PAM:setcred grantors=pam_securetty,pam_shells,pam_faillock,pam_permit,pam_faillock acct="mint" exe="/usr/bin/login" hostnam> Apr 08 15:59:18 Indiana audit[4154]: USER_START pid=4154 uid=0 auid=1000 ses=12 msg='op=PAM:session_open grantors=pam_loginuid,pam_keyinit,pam_limits,pam_unix,pam_permit,pam_mail,pam_systemd,pam_env acct="mint"> Apr 08 15:59:18 Indiana login[4154]: pam_env(login:session): deprecated reading of user environment enabled (In reply to mintjulep from comment #41) > As for the 2nd solution, I don't know how to create an executable file like > you did No problemo! The steps are something like this: Create and edit the custom file: sudo nano /etc/X11/xinit/xinitrc.d/99-reset-failed-user-service.sh Fill it with these two lines: #!/bin/sh systemctl --user reset-failed Now save and exit (with CTRL + X and choosing "y" if using nano text editor) Now give it executable permission: sudo chmod +x /etc/X11/xinit/xinitrc.d/99-reset-failed-user-service.sh In the future, you can simply delete this custom file, as it's no longer needed if the updated KDE packages resolve this issue: sudo rm /etc/X11/xinit/xinitrc.d/99-reset-failed-user-service.sh Hope that helps. :) Ok ! followed your instructions (thank you very much by the way, I relay need to learn about shell scripts) Tryed to logout login just after reboot, worked fine. Tryed to run some app (firefox) and logout login, worked fine. Tryed to run several app, wait for a couple minute with normal usage, logout, login and... It worked too ! Seems like a resolved bug to me !! Keep you posted if I have any more issue. Thank you very much to everybody who helped on this. This is not a very important usecase (less than ever as with an ssd disk, complete reboot is incredibly fast) but it has been bugging me for month to have this little malfunction that I couldn't solve. *** Bug 429415 has been marked as a duplicate of this bug. *** *** Bug 434369 has been marked as a duplicate of this bug. *** |