gnome-shell crashes with SIGABRT: mutter:ERROR:backends/meta-monitor-manager.c:2274:meta_monitor_manager_get_logical_monitor_from_number: assertion failed: ((unsigned int) number < g_list_length (manager->logical_monitors))

Sounds like you might have hit bug 1714330 but we should go though the proper procedure in case you're experiencing a different problem...

Thank you for taking the time to report this bug and helping to make Ubuntu better. However, your crash report is either missing or challenging to deal with as a ".crash" file. Please follow these instructions to have apport report a new bug about your crash that can be dealt with by the automatic retracer.

If you are running the Ubuntu Stable Release you might need to enable apport in /etc/default/apport and restart.

Now open your file manager, navigate to your /var/crash directory and open the crash report you wish to submit.
If this fails you will have to open a terminal and file your report with 'ubuntu-bug /var/crash/_my_crash_report.crash' where _my_crash_report.crash is the crash you would like to report. If you get an error that you aren't allowed to access this report you will have to file it with 'sudo ubuntu-bug /var/crash/_my_crash_report.crash'.

I'm closing this bug report since the process outlined above will automatically open a new bug report which can then dealt with more efficiently. Thanks in advance for your cooperation and understanding.

Apport is enabled in /etc/default/apport but no file is located /var/crash even after crash. I have created the bug report using ubuntu-bug tool.

Please how to force .crash file creation?

Have a look in your /var/crash/

If there are none there already then I'm not sure why, other than dumping of core somehow failed, or less likely it's not "crashing" via any of the normal fatal signals.

Hello. /var/crash/ is empty. I have attached part of Syslog starting at the time I switched off the monitor and a lot of errors were logged since that time. Please look at it, may be it can help also determine why crash file is not generated.

Thanks. Your log shows gnome-shell is crashing in libmutter:

Sep 13 14:08:10 internet-ws kernel: [11508.541040] gnome-shell[3056]: segfault at 2c ip 00007fd7c9d0cb24 sp 00007ffc4ffe2a28 error 4 in libmutter-1.so.0.0.0[7fd7c9c6f000+13f000]
Sep 13 14:08:10 internet-ws gnome-session[2964]: gnome-session-binary[2964]: WARNING: Application 'org.gnome.Shell.desktop' killed by signal 11
Sep 13 14:08:10 internet-ws gnome-session-binary[2964]: WARNING: Application 'org.gnome.Shell.desktop' killed by signal 11

So it does sound like this might be bug 1714330. Please fully update your system and tell us if the problem persists.

The system is up to date. Unfortunately the problem persist.
I'm Sorry, I have added old syslog file in previous post. Please find attached syslog from today morning, logs starting on the time of crash.

Aha!

Yes you were probably experiencing bug 1714330 in the old log of comment #6 but not any more.

In your new log of comment #8 you are now experiencing a new assertion:

Sep 18 07:25:39 internet-ws org.gnome.Shell.desktop[1891]: mutter:ERROR:backends/meta-monitor-manager.c:2274:meta_monitor_manager_get_logical_monitor_from_number: assertion failed: ((unsigned int) number < g_list_length (manager->logical_monitors))

... preceded by Javascript errors about the primary monitor.

Please try deleting this file in your home directory:

  ~/.config/monitors.xml

and then reboot.

hmm, no such file ~/.config/monitors.xml in my home directory

I mean "~" is your home directory, ".config" is a directory and "monitors.xml" is the file. :)

Regardless, your original description suggests the invalid config occurs dynamically. Not necessarily on disk.

Found it. Gnome Developers are already aware of the crash:
https://bugzilla.gnome.org/show_bug.cgi?id=787637

If you would like to contribute to a speedy resolution then please comment on that page.

Thank you, yes It seems that the bug report you found on Gnome Bugzilla is reporting the same type of crash.

Today the update of mutter to version 3.26.1-1 was installed on my PC by "apt upgrade" and the issue still persist. Please see attached syslog, cut from the time of crash (caused by monitor switch off).

Also tracking in:
https://errors.ubuntu.com/problem/a9ace2e3301009e3a5344f8dac4e56fc4c23b1b8

may be the status of the bag should be changed from "Fix Released" to "Confirmed". Related GNOME bug #787637 is reopened.

We can't change upstream bug link statuses explicitly but removing the upstream task and recreating it will hopefully do the trick.

It did.Thank you Daniel.

The crash itself is fixed by https://github.com/GNOME/mutter/commit/f044511037a51ae9c438addf3a8374f559916aae (part of fixing bug #1722811), so probably we can join the two bugs together for the ubuntu side.

The root cause of this is probably in a different spot though.
Maybe related to https://bugzilla.gnome.org/show_bug.cgi?id=788607

I confirm that this crash is fixed by https://github.com/GNOME/gnome-shell/commit/db43c45b12fbae462857e59dce6849c6aa965091

This bug was fixed in the package gnome-shell - 3.26.1-0ubuntu3

---------------
gnome-shell (3.26.1-0ubuntu3) artful; urgency=medium

  * Cherry-pick git_layout-unset-when-headless.patch from gnome-3-26
    branch to fix crash bug (LP: #1717170)

 -- Jeremy Bicha <email address hidden> Fri, 13 Oct 2017 16:10:59 -0400

I had one crash after applying this yesterday (and rebooting): it crashed right out to the console, not even to the login screen, and when I tried logging in again the session failed to start. But since a second reboot it's apparently been fine. This bug went away, and so too has my bug #1720149 (so far anyway, I don't seem able to reproduce it now, although saying so will doubtless trigger it next time I let the displays sleep). I can't explain that first crash, but as I also can't repeat it, I guess we ignore it for now.

This in Gnome in Xorg under nVidia 387, modeset=0. xorg sessions seem unable to start with modeset=1 btw.

bug #1723615 came up with this fix in place. Whack-a-mole! :-)

Adding this to *this* bug, despite the "Fix Released", because it exactly matches the original post for this bug, whereas the others I've encountered and reported on are different:

Oct 14 20:02:45 fleetfoot gnome-shell[12750]: meta_monitor_manager_get_logical_monitor_from_number: assertion '(unsigned int) number < g_list_length (manager->logical_monitors)' failed
Oct 14 20:02:45 fleetfoot gnome-shell[12750]: meta_workspace_get_work_area_for_monitor: assertion 'logical_monitor != NULL' failed
Oct 14 20:02:45 fleetfoot kernel: [29646.218484] gnome-shell[12750]: segfault at 40 ip 00007f9354b11e60 sp 00007ffd0b949810 error 4 in libmutter-1.so.0.0.0[7f9354a6c000+141000]
Oct 14 20:02:46 fleetfoot gnome-shell[25511]: Couldn't register with accessibility bus: 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.
Oct 14 20:02:46 fleetfoot gnome-shell[25511]: JS WARNING: [resource:///org/gnome/shell/ui/main.js 315]: reference to undefined property "MetaStage"
Oct 14 20:02:46 fleetfoot gnome-shell[25511]: JS WARNING: [resource:///org/gnome/shell/ui/layout.js 221]: reference to undefined property "MetaWindowGroup"
Oct 14 20:02:46 fleetfoot gnome-shell[25511]: JS WARNING: [resource:///org/gnome/shell/ui/osdMonitorLabeler.js 59]: reference to undefined property "MetaDBusDisplayConfigSkeleton"
Oct 14 20:02:46 fleetfoot gnome-shell[25511]: JS WARNING: [resource:///org/gnome/gjs/modules/tweener/tweener.js 540]: reference to undefined property "isSpecialProperty"
Oct 14 20:02:46 fleetfoot gnome-shell[25511]: JS WARNING: [resource:///org/gnome/shell/ui/slider.js 38]: reference to undefined property "CallyActor"
Oct 14 20:02:46 fleetfoot gnome-shell[25511]: meta_monitor_manager_get_logical_monitor_from_number: assertion '(unsigned int) number < g_list_length (manager->logical_monitors)' failed
Oct 14 20:02:46 fleetfoot gnome-shell[25511]: meta_workspace_get_work_area_for_monitor: assertion 'logical_monitor != NULL' failed
Oct 14 20:02:46 fleetfoot gnome-shell[25511]: Extension "<email address hidden>" had error: TypeError: this._monitor is undefined
Oct 14 20:02:46 fleetfoot gnome-shell[25511]: loading user theme: /usr/share//themes/Arc-Dark/gnome-shell/gnome-shell.css
Oct 14 20:02:46 fleetfoot gnome-shell[25511]: Some code accessed the property 'DBusMenu' on the module 'interfaces'. That property was defined with 'let' or 'const' inside the module. This was previously supported, but is not correct according to the ES6 standard. Any symbols to be exported from a module must be defined with 'var'. The property access will work as previously for the time being, but please fix your code anyway.
Oct 14 20:02:46 fleetfoot gnome-shell[25511]: Some code accessed the property 'WORKAROUND_RELOAD_TYPE_REGISTER' on the module 'util'. That property was defined with 'let' or 'const' inside the module. This was previously supported, but is not correct according to the ES6 standard. Any symbols to be exported from a module must be defined with 'var'. The property access will work as previously for the time being, ...

What I was originally reported is not fixed. It means, gnome-shell session crash when the monitor is disconnected/switch off or its input is changed.
Mutter and Gnome-Shell updated to:
mutter: 3.26.1-2ubuntu1
gnome-shell: 3.26.1-0ubuntu3

Ryzko,

Do you have a stacktrace for the crash you're getting now?
The one related to this title is definitely fixed, and errors.ubuntu.com confirms it.

What you're getting might be more related to https://pad.lv/1723048 ?

Well, only what I have is again part of syslog (attached). If you can advice me how to provide better logs, it would be appreciated. However, it seems that the crash now is not caused by monitor switch off but by switching it on again, according to syslog. Nothing is logged between switch off and switch on.

I was just retested the crash and it goes worse, now it is complete GUI crash not a "just session kick-out". I have created new bug report 1724557 due the fact that this bug 1717170 has fix released status and also because the behaviour and logs changed.

Now when I switch my display back on again, I see my desktop and open windows for a short while, and then am immediately kicked out of the session. Before the entire desktop would die, now I'm just back at user select screen.

On two separate systems.

This bug is closed, rightly or wrongly. It would help us if all further discussion was in different bugs that are open. Even if it feels like the same issue as was discussed here...

I have the same crash as Rachel whenever I unplug the DVI from my NVidia graphics card or switch the KVM to another monitor. I use Ubuntu 17.10 latest, with mutter 3.26.2-0ubuntu0.1 installed. This does not fix the crash.

I also get the assertion: meta_monitor_manager_get_logical_monitor_from_number: assertion '(unsigned int) number < g_list_length (manager->logical_monitors)' failed

I could fix the issue by building my own mutter based on 3.26.2-0ubuntu0.1 but with the patch 3572502 from https://git.gnome.org/browse/mutter/commit/?id=3572502 applied. The patch applies cleanly and the crash is gone.

Please instruct where I can report this, since this bug here is closed?

Mario, ideally please open a new bug by running 'ubuntu-bug mutter'. Or if that doesn't work then go here:
https://bugs.launchpad.net/ubuntu/+source/mutter/+filebug