Bug 1431618

Summary: anaconda aborts (Pane is dead) with GLib-GObject:ERROR:gsignal.c:693:handler_unref_R: assertion failed: (hlist->tail_after == handler)
Product: Red Hat Enterprise Linux 7 Reporter: Radek Vykydal <rvykydal>
Component: anacondaAssignee: Anaconda Maintenance Team <anaconda-maint-list>
Status: CLOSED WONTFIX QA Contact: Release Test Team <release-test-team-automation>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 7.3CC: klember, mclasen
Target Milestone: rc   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of:
: 1432436 (view as bug list) Environment:
Last Closed: 2021-01-15 07:32:47 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 1432436    
Attachments:
Description Flags
backtrace from core dump
none
program.log
none
anaconda.log
none
syslog none

Description Radek Vykydal 2017-03-13 12:41:52 UTC 
Created attachment 1262431 [details]
backtrace from core dump

Happens during 7.4 devel phase, eg with RHEL-7.4-20170310.n.0 compose.

When running anaconda kickstarts tests which run 4 installations in virtual machines in parallel on kvm hypervisor we are sometimes [1] hitting this crash:

GLib-GObject:ERROR:gsignal.c:693:handler_unref_R: assertion failed: (hlist->tail_after == handler)

Seems to be hit by this installer function in widgets/src/BaseWindow.c [2]:

void anaconda_base_window_clear_info(AnacondaBaseWindow *win) {
    if (!win->priv->info_shown)
        return;

    gtk_widget_hide(win->priv->info_bar);
->    gtk_widget_destroy(win->priv->info_bar);
    gtk_widget_destroy(win->priv->event_box);
    win->priv->info_shown = FALSE;
}


I am attaching backtrace.

NOTE: Some time ago we have been hitting also
GLib:ERROR:gmain.c:3141:g_main_dispatch: assertion failed: (source->context == context) 
in some installation instances but I've not been able to hit it last few weeks.

[1] It can be hit about every second (almost every) kickstart test batch running 11 tests (kickstart installations). The tests are scheduled so that 4 tests are run in parallel on the hypervisor.

[2] The code has been reworked on master:
https://github.com/rhinstaller/anaconda/commit/5b4513c104ff7fc3d04c469abcc997320f8827ef
Maybe it would fix the issue but there seems to be a wider context of the change.
Comment 2 Radek Vykydal 2017-03-13 12:43:47 UTC 
Created attachment 1262432 [details]
program.log
Comment 3 Radek Vykydal 2017-03-13 12:44:45 UTC 
Created attachment 1262433 [details]
anaconda.log
Comment 4 Radek Vykydal 2017-03-13 12:45:13 UTC 
Created attachment 1262434 [details]
syslog
Comment 5 Radek Vykydal 2017-03-13 12:47:37 UTC 
(In reply to Radek Vykydal from comment #0)

> [2] The code has been reworked on master:
> https://github.com/rhinstaller/anaconda/commit/
> 5b4513c104ff7fc3d04c469abcc997320f8827ef
> Maybe it would fix the issue but there seems to be a wider context of the
> change.

David, you've been working on this stuff in Anaconda, do you have any hint?
Comment 6 David Shea 2017-03-13 14:44:17 UTC 
The change in 5b4513c1 was just to add the animation, not to address any particular bug. Based on the backtrace, it looks like either info_bar is being destroyed more than once, or there is a bug in gtk in the handling of the signal list.
Comment 7 Radek Vykydal 2017-03-14 12:54:14 UTC 
I can't see anything wrong on installer side.
Kalev (or someone from Gtk/Glib team), could you please check the backtrace in the description (Thread1) if it looks fine for you?
Are there any known issues or hints for debugging or pitfalls regarding the failing assert would point to?
Thank you.
Comment 8 Kalev Lember 2017-03-15 09:39:50 UTC 
Not sure what's going on, but maybe running anaconda under valgrind would reveal something? Adding mclasen to CC in case he has ideas.

Also, we have new glib2 and gtk3 coming to 7.4 (https://errata.devel.redhat.com/advisory/27092), would be interesting to know if this changes anything with this crash.
Comment 9 Radek Vykydal 2017-03-30 09:40:20 UTC 
Unlike the bug 1432436, I haven't hit the issue after the glib2 and gtk3 rebase so far.
Comment 11 RHEL Program Management 2021-01-15 07:32:47 UTC 
After evaluating this issue, there are no plans to address it further or fix it in an upcoming release.  Therefore, it is being closed.  If plans change such that this issue will be fixed in an upcoming release, then the bug can be reopened.