update-manager crashed with OSError in _execute_child()

Bug #244739 reported by Derrick Hudson
6
Affects Status Importance Assigned to Milestone
update-manager (Ubuntu)
Triaged
Medium
Unassigned

Bug Description

Binary package hint: update-manager

update-manager needs to catch and handle failure to execute gksu. At the time of this crash, gksu was not executable. My report is not about execute permission on gksu, but rather about update-manager handling the error more gracefully.

ProblemType: Crash
Architecture: i386
Date: Tue Jul 1 21:08:15 2008
DistroRelease: Ubuntu 8.04
ExecutablePath: /usr/bin/update-manager
InterpreterPath: /usr/bin/python2.5
Package: update-manager 1:0.87.24
PackageArchitecture: all
ProcCmdline: /usr/bin/python2.5 /usr/bin/update-manager
ProcEnviron:
 PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/bin/X11:/usr/games
 LANG=en_US.UTF-8
 SHELL=/bin/zsh
PythonArgs: ['/usr/bin/update-manager']
SourcePackage: update-manager
Title: update-manager crashed with OSError in _execute_child()
Uname: Linux 2.6.24-18-generic i686
UserGroups: adm admin audio cdrom dialout floppy fuse lpadmin plugdev scanner src staff video

Tags: apport-crash
Revision history for this message
Derrick Hudson (dman) wrote :
Revision history for this message
Michael Vogt (mvo) wrote :

Thanks,

this happens when /tmp is mounted nonexec usually. Could you please check/confirm this?

Changed in update-manager:
importance: Undecided → Medium
milestone: none → intrepid-alpha-3
status: New → Triaged
Revision history for this message
Derrick Hudson (dman) wrote :

No, this crash was not caused by mount flags on /tmp. It was caused by removing execute permission from the gksu binary in /usr/bin. (I removed that permission bit, not some part of the Ubuntu system) The stack trace gathered by the crash reporter tool shows a different starting point than the one in #219518.

Only tangential to this, gksu usually fails to display the first time update-manager executes it. After I kill it, then retry the upgrade it works. I was experimenting to try find a resolution. Since I couldn't uninstall gksu, I removed the execute bit expecting udpate-manager to skip it and try gnomesu instead. That is how I arrived at this situation.

Although update-manager is not responsible for my interference with the system, it shouldn't crash. It should catch the error and, at the very least, simply display an error message.

Unless the patch you uploaded corrects run_synaptic() in UpdateManager.py, this bug should not be marked a duplicate of #219518.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.