[Bugs] [Bug 243] New: proposed patch to fix system hang on shutdown with psmisc-22.8 and gnome 2.28.0

bugzilla-daemon at exherbo.org bugzilla-daemon at exherbo.org
Tue Nov 3 05:15:53 GMT 2009


https://bugs.exherbo.org/show_bug.cgi?id=243

           Summary: proposed patch to fix system hang on shutdown with
                    psmisc-22.8 and gnome 2.28.0
           Product: Repositories
           Version: unspecified
          Platform: All
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P5
         Component: Packages
        AssignedTo: unassigned at exherbo.org
        ReportedBy: li.erik at gmail.com
           Package: psmisc-22.8


Created an attachment (id=140)
 --> (https://bugs.exherbo.org/attachment.cgi?id=140)
patch to fix process killing behavior

After a recent package upgrade (including a gnome upgrade), I noticed that my
system would hang at "unmounting filesystems" during a shutdown. This would
only happen after logging in at least once with gdm into a gnome session. I am
running gnome 2.28.0.

After looking through /etc/init.d/halt.sh and adding some debug statements, I
found that /home was not getting unmounted because gvfsd-metadata was not
getting killed and was accessing the /home partition.

However, the halt.sh script already was using

#fuser -s -k -9 -m /home

in order to take care of this sort of problem. Fuser is part of the psmisc
package and was not properly sending the kill signal to the processes it
identifies. If you run fuser in interactive mode, you can see that the process
tries to call kill on pid 0, which effectively terminates the fuser process
itself. The proposed patch fixes this behavior and my computer shutdowns
normally.

Is this potentially also a problem with gvfsd-metadata? The fuser patch
produces the correct behavior of the system overall, but I was wondering
whether or not gvfsd-metadata should hang around so long anyways.


-- 
Configure bugmail: https://bugs.exherbo.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.



More information about the Bugs mailing list