| Summary: | Can't log out of Plasma 6 | ||
|---|---|---|---|
| Product: | [Plasma] plasmashell | Reporter: | Synthetic451 <bugs.kde.org.facelift226> |
| Component: | Session Management | Assignee: | Plasma Bugs List <plasma-bugs> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | danilo.luvizotto, mike.cloaked, mss, omarhanykasban706, sitter, web |
| Priority: | NOR | ||
| Version: | 6.0.3 | ||
| Target Milestone: | 1.0 | ||
| Platform: | Arch Linux | ||
| OS: | Linux | ||
| See Also: | https://bugs.kde.org/show_bug.cgi?id=484864 | ||
| Latest Commit: | Version Fixed In: | ||
| Sentry Crash Report: | |||
| Attachments: |
logs after triggering a logout
journal that includes lines for logout fail from plasma 6 |
||
|
Description
Synthetic451
2024-04-04 05:28:06 UTC
*** Bug 485129 has been marked as a duplicate of this bug. *** I see the same issue. Looks like a deadlock in systemd caused by drkonqi-coredump-processor. This started after an update of KDE Neon from 6.0.2 to 6.03. Apr 07 19:22:07 xxx plasmashell[1886]: file:///usr/share/plasma/plasmoids/org.kde.plasma.kickoff/contents/ui/ApplicationsPage.qml:202: TypeError: Cannot call method 'forceActiveFocus' of null Apr 07 19:22:10 xxx dbus-daemon[1269]: [session uid=1001 pid=1269] Activating service name='org.kde.LogoutPrompt' requested by ':1.27' (uid=1001 pid=1886 comm="/usr/bin/plasmashell --no-respawn " label="unconfined") Apr 07 19:22:10 xxx dbus-daemon[1269]: [session uid=1001 pid=1269] Successfully activated service 'org.kde.LogoutPrompt' Apr 07 19:22:11 xxx ksmserver-logout-greeter[2238]: qt.gui.imageio: libpng warning: iCCP: known incorrect sRGB profile Apr 07 19:22:11 xxx ksmserver-logout-greeter[2238]: kf.windowsystem: static bool KX11Extras::compositingActive() may only be used on X11 Apr 07 19:22:13 xxx dbus-daemon[1269]: [session uid=1001 pid=1269] Activating service name='org.kde.Shutdown' requested by ':1.48' (uid=1001 pid=2238 comm="/usr/lib/x86_64-linux-gnu/libexec/ksmserver-logout" label="unconfined") Apr 07 19:22:13 xxx dbus-daemon[1269]: [session uid=1001 pid=1269] Successfully activated service 'org.kde.Shutdown' Apr 07 19:22:13 xxx kded6[1863]: Service ":1.43" unregistered Apr 07 19:22:13 xxx systemd[1222]: Stopped target plasma-workspace-wayland.target. Apr 07 19:22:13 xxx systemd[1222]: Stopped target KDE Plasma Workspace. Apr 07 19:22:13 xxx systemd[1222]: Requested transaction contradicts existing jobs: Transaction for graphical-session.target/stop is destructive (drkonqi-coredump-pickup.service has 'start' job queued, but 'stop' is included in transaction). Apr 07 19:22:13 xxx systemd[1222]: graphical-session.target: Failed to enqueue stop job, ignoring: Transaction for graphical-session.target/stop is destructive (drkonqi-coredump-pickup.service has 'start' job queued, but 'stop' is included in transaction). Apr 07 19:22:13 xxx systemd[1222]: Stopped target KDE Plasma Workspace Core. Apr 07 19:22:13 xxx systemd[1222]: Stopped target Startup of XDG autostart applications. Apr 07 19:22:13 xxx systemd[1222]: Stopped target Session services which should run early before the graphical session is brought up. Operating System: KDE neon 6.0 KDE Plasma Version: 6.0.3 KDE Frameworks Version: 6.0.0 Qt Version: 6.6.3 Kernel Version: 6.5.0-26-generic (64-bit) Graphics Platform: Wayland Processors: 4 × Intel® Core™ i5-6300U CPU @ 2.40GHz Memory: 7,6 GiB of RAM Graphics Processor: Mesa Intel® HD Graphics 520 Manufacturer: LENOVO Product Name: 20F5S1H800 System Version: ThinkPad X260 Yes, it seems like drkonqi is the culprit here. If I uninstall it (along with plasma-meta which depends on it), then log outs work fine every time. Reinstalling drkonqi immediately causes the issue to come back. I can also confirm that uninstalling drkonqi and plasma-meta leads to the logout problem no longer occurring. Hopefully this will allow the underlying cause of this bug to be fixed so that installing fixed a fixed version of drkonqi will allow logout to complete without problems again. Instead of uninstalling drkonqi these command should be enough to work around the issue: systemctl --user stop drkonqi-coredump-pickup.service systemctl --user mask drkonqi-coredump-pickup.service Don't forget to execute this once the issue is fixed: systemctl --user unmask drkonqi-coredump-pickup.service Should this bug be reassigned to drkonqi or are the relevant people already aware of this? I added Harald Sitter to the CC list but am not really sure if DrKonqi is really the root cause or just triggering another systemd dependency issue. Why are the other jobs stopped (eg. "Stopped target KDE Plasma Workspace.") even though something is wrong and do not really end the session and thus throw the whole session management into disorder? Some oddities I see here: 1. This affects only one of my two Notebooks running Neon 6.0.3 2. On the affected one only one user is affected. I cans till switch the session to another user and restart the system normally. 3. It is also not reliably affecting that user, maybe 80% of the cases. 4. The first time this happens I see the logout confirmation screen (I guess that is org.kde.LogoutPrompt) but when I confirm there nothing happens (I guess that should be handled by org.kde.Shutdown). 5. Once this happened once I don't ever see that confirmation screen for this user again and the logout button just doesn't do anything. This is https://bugs.kde.org/show_bug.cgi?id=484864 isnt it? Looks very much like that one, yes. Shall we mark this as a duplicate or is there a deeper issue in the session handling itself which should be looked into? Couldn't other units break this as well? I found this: > systemd has a minimal transaction system: if a unit is requested to start up or shut down it will add it and all its dependencies to a temporary transaction. Then, it will verify if the transaction is consistent (i.e. whether the ordering of all units is cycle-free). If it is not, systemd will try to fix it up, and removes non-essential jobs from the transaction that might remove the loop. Also, systemd tries to suppress non-essential jobs in the transaction that would stop a running service. Finally it is checked whether the jobs of the transaction contradict jobs that have already been queued, and optionally the transaction is aborted then. If all worked out and the transaction is consistent and minimized in its impact it is merged with all already outstanding jobs and added to the run queue. Effectively this means that before executing a requested operation, systemd will verify that it makes sense, fixing it if possible, and only failing if it really cannot work. Source: https://www.freedesktop.org/software/systemd/man/latest/systemd.html The root cause here seems to be that all units are transitioned to stopped except graphical-session which makes the whole session more or less a zombie. Maybe if the stopping of graphical-session could be marked as an essential job somehow the whole transaction would fail which should be a better outcome than this. systemctl has a --job-mode switch for this and the shutdown/halt commands set the equivalent to "replace-irreversibly" according to the man page. I suppose we should at least surface an error in the UI. I digged a bit and it looks like the job is already run in "fail" mode but unless I miss something then the result of the returned job object is never checked: https://invent.kde.org/plasma/plasma-workspace/-/blob/v6.0.3/startkde/plasma-shutdown/shutdown.cpp?ref_type=tags#L126 There should probably some error message here, yes. Created attachment 168282 [details]
journal that includes lines for logout fail from plasma 6
Journal log including logout fail lines prior to masking the drkonqi-coredump-pickup.service
This seems to have been fixed with Plasma 6.0.4 https://invent.kde.org/plasma/drkonqi/-/commit/45fcd31c993cfa518e5957f4dcff51ffeb0d7405 https://bugs.kde.org/show_bug.cgi?id=484864 I can no longer reproduce this. Closing. i get `ksmserver-logout-greeter[5072]: kf.windowsystem: static bool KX11Extras::compositingActive() may only be used on X11` on 6.2.3 please reopen it its not fixed (In reply to Omar Hany Kasban from comment #16) > i get `ksmserver-logout-greeter[5072]: kf.windowsystem: static bool > KX11Extras::compositingActive() may only be used on X11` on 6.2.3 > please reopen it its not fixed masking drkonqi / removing drkonqi, doesn't work this time (In reply to Omar Hany Kasban from comment #17) > (In reply to Omar Hany Kasban from comment #16) > > i get `ksmserver-logout-greeter[5072]: kf.windowsystem: static bool > > KX11Extras::compositingActive() may only be used on X11` on 6.2.3 > > please reopen it its not fixed > > masking drkonqi / removing drkonqi, doesn't work this time This looks like a different/new bug. Can you please report a new one? fixed it was lan-mouse (In reply to Malte S. Stretz from comment #18) > (In reply to Omar Hany Kasban from comment #17) > > (In reply to Omar Hany Kasban from comment #16) > > > i get `ksmserver-logout-greeter[5072]: kf.windowsystem: static bool > > > KX11Extras::compositingActive() may only be used on X11` on 6.2.3 > > > please reopen it its not fixed > > > > masking drkonqi / removing drkonqi, doesn't work this time > > This looks like a different/new bug. Can you please report a new one? https://github.com/feschber/lan-mouse/issues/237 it was lan-mouse problem |