Bug 699921
| Summary: | gdm-greeting-simple using 100% CPU | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Jason Haar <jhaar> | ||||||||
| Component: | gdm | Assignee: | jmccann | ||||||||
| Status: | CLOSED DUPLICATE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> | ||||||||
| Severity: | medium | Docs Contact: | |||||||||
| Priority: | unspecified | ||||||||||
| Version: | 15 | CC: | cschalle, jmccann, rstrode | ||||||||
| Target Milestone: | --- | ||||||||||
| Target Release: | --- | ||||||||||
| Hardware: | x86_64 | ||||||||||
| OS: | Linux | ||||||||||
| Whiteboard: | |||||||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||||||
| Doc Text: | Story Points: | --- | |||||||||
| Clone Of: | Environment: | ||||||||||
| Last Closed: | 2011-05-11 17:45:35 UTC | Type: | --- | ||||||||
| Regression: | --- | Mount Type: | --- | ||||||||
| Documentation: | --- | CRM: | |||||||||
| Verified Versions: | Category: | --- | |||||||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||||||
| Embargoed: | |||||||||||
| Attachments: |
|
||||||||||
|
Description
Jason Haar
2011-04-26 23:01:14 UTC
I think I found a pattern in syslog When I enter "domain\username", it triggers the following Apr 27 12:32:24 jhaar-fc-nzcl gdm-simple-greeter[1436]: Gtk-WARNING: gtkwidget.c:6778: widget not within a GtkWindow Apr 27 12:32:24 jhaar-fc-nzcl gdm-simple-greeter[1436]: Gtk-WARNING: gtk_widget_size_allocate(): attempt to allocate widget with width -47 and height -47 Apr 27 12:32:24 jhaar-fc-nzcl dbus-daemon: [system] Activating via systemd service name='org.freedesktop.Accounts' unit='accounts-daemon.service' This is then immediately followed by a gdm "crash" (the screen doesn't restart, so it couldn't be a true crash?). This is then IMMEDIATELY followed by a crash of NetworkManager. I don't know if they are related, other than I can trigger them at will by trying to login. I can also fix it by killing gdm-greeting-simple and trying again I'll attach the logs. Created attachment 495153 [details]
gdm-crash.log
Created attachment 495154 [details]
NetworkManager-crash
I've noticed another issue - maybe different. This is a new system, so I'm obviously installing software all over the place. Whenever I install software that installs a "systemd" based service, gdm crashes and I end up thrown back to the login prompt. Attached is the gdm crash details from syslog for the latest. Even running systemctl commands causes it :-( Created attachment 495312 [details]
systemd-gdm-crash
FYI I noticed updates for gnome were pushed out, so I upgraded. Now the way it fails is different - but it looks like the root cause is the same I enter "domain\username" and it prompts for my password. This time "Password" prompt actually works and it accepts my password and states "the domain is not available, blah offline mode" (sorry, can't remember the exact string). That window sits there and never exits :-( As it says, at that moment I'd be logging in via winbind's "offline" mode. When I Ctrl-Alt-F2 and do a "ps", I can see gdm-greeting-simple has called "pam: gdm-password". Maybe the fact winbind outputs the "offline mode" comment has confused it? If I kill gdm-password, nothing changes, but - as before - killing gdm-greeting-simple does the trick. I end up back at the gdm login prompt staring at "Jason Haar" instead of "Other" and typing my password works: no "offline mode" popup - just straight in. Here's a thought: is the "offline mode" popup coded to only occur the first time a user logs in, and that popup (or the cause of the popup) is the cause of all this? If so, can pam be told not to stderr/stdout that error (it could do it via syslog?). Or, can gdm be told to ignore that error string from pam? Windows doesn't bother telling the user when they're in offline mode during login - so why should Unix? I may be totally on the wrong path there of course ;-) This should be fixed with very recent versions of account service i think. *** This bug has been marked as a duplicate of bug 703569 *** |