| Summary: | Shutdown/reboot from kmserver-logout-greeter makes re-opened apps think they crashed | ||
|---|---|---|---|
| Product: | [Plasma] plasmashell | Reporter: | edgars.polis |
| Component: | Session Management | Assignee: | David Edmundson <kde> |
| Status: | RESOLVED WORKSFORME | ||
| Severity: | normal | CC: | nate, plasma-bugs-null |
| Priority: | NOR | ||
| Version First Reported In: | 5.24.0 | ||
| Target Milestone: | 1.0 | ||
| Platform: | unspecified | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed In: | ||
| Sentry Crash Report: | |||
| Attachments: |
Cat of ksmserverrc
ksmserverrc file after a reboot |
||
|
Description
edgars.polis
2021-04-03 15:19:35 UTC
We are respecting the setting, as nothing is happening until you manually launch the browser. ksmserver startup sounds correct. Can you maybe upload your ~/.config/ksmserverrc and/or verify there is no section titled: [Session: saved at previous logout] What we're seeing here is clients doing something completely internal to the client. What is interesting is that they're claiming they crashed. Can you read the end of 'journalctl --user -b -1' (user logs from last boot) and see if it mentions these apps. Do we see something similar with dolphin/kate? Created attachment 137313 [details]
Cat of ksmserverrc
A cat output of ~/.config/ksmserverrc next to a crash recovery brave browser and setting indicating "start with an empty session"
Created attachment 137314 [details]
ksmserverrc file after a reboot
ksmserverrc file from my VM after a reboot with Brave browser open
> Can you read the end of 'journalctl --user -b -1' > (user logs from last boot) and see if it mentions these apps. I do not, I even tried `journalctl --user -b -1 | grep firefox` and it returned no results > Do we see something similar with dolphin/kate? I could not reproduce it with kate. If I save it, then there is no issue If I do not, then it will postpone the logout until I take action One interesting thing that I notice, is that these windows "stick around" for a while when even the plasma shell seems to be stopped. As far as I can tell they are still fully interactive. You can clearly see that here https://youtu.be/smId0VmppC0?t=137 (timestamp included) It may be that they are behaving unexpectedly and not closing the way plasma wants, so they are terminated. Rebooting with both Kate and Firefox open, Kate closes immediately, but Firefox can still be seen for a good 3 seconds surrounded by a blank screen > One interesting thing that I notice, is that these windows "stick around"
> for a while when even the plasma shell seems to be stopped. As far as I can
> tell they are still fully interactive.
> You can clearly see that here https://youtu.be/smId0VmppC0?t=137 (timestamp
> included)
>
> It may be that they are behaving unexpectedly and not closing the way plasma
> wants, so they are terminated.
> Rebooting with both Kate and Firefox open, Kate closes immediately, but
> Firefox can still be seen for a good 3 seconds surrounded by a blank screen
Sorry, I meant Brave browser. Firefox seems to close immediately
I ran `brave --enable-logging=stderr --v=1 > log.txt 2>&1` to open the Brave browser with logging enabled, and rebooted the system both with the prompt and with a command. First thing that I have noticed is that rebooting with a command(no crashes) is that the log file includes a "Histogram" And the second is that rebooting with the prompt (Brave thinks that it as crashed) the last two log entries are: > [2211:2211:0403/125901.829967:ERROR:connection.cc(66)] X connection error received. > [2211:2227:0403/125901.829900:ERROR:connection.cc(66)] X connection error received. Another time it is: > [1989:1989:0403/131034.610368:WARNING:browser_main_loop.cc(273)] Gdk: brave: Fatal IO error 11 (Resource temporarily unavailable) on X server :0. root And another: > [1825:1831:0403/131229.024627:ERROR:connection.cc(66)] X connection error received. > [1649:1649:0403/131229.025664:WARNING:browser_main_loop.cc(273)] Gdk: brave: Fatal IO error 11 (Resource temporarily unavailable) on X server :0. > [1825:1825:0403/131229.025584:ERROR:connection.cc(66)] X connection error received. So it may be that the X server gets shut down before the browser has the chance to close? That may explain why it has no problems when using `shutdown -r now`, because they probably kill processes in a different order Is this still happening in a recent version of Plasma, like 6.2.5 or later? 🐛🧹 ⚠️ This bug has been in NEEDSINFO status with no change for at least 15 days. Please provide the requested information, then set the bug status to REPORTED. If there is no change for at least 30 days, it will be automatically closed as RESOLVED WORKSFORME. For more information about our bug triaging procedures, please read https://community.kde.org/Guidelines_and_HOWTOs/Bug_triaging. Thank you for helping us make KDE software even better for everyone! 🐛🧹 This bug has been in NEEDSINFO status with no change for at least 30 days. Closing as RESOLVED WORKSFORME. |