[Pkg-utopia-maintainers] Bug#561505: hal: Fails to upgrade

[Kurt Roeckx]

On Thu, Dec 17, 2009 at 09:30:48PM +0100, Michael Biebl wrote:
> Kurt Roeckx wrote:
> > I was just doing a dist-upgrade from lenny to the current squeeze.
> > hal is failing to upgrade.  I'm not really sure what happened
> > exactly, since it scrolled of my screen.
> > 
> > I then tried to do "apt-get -f install", hoping that everything
> > would be fixed, or that I would get the same error message again.
> > But it seems some hald was started before, and the postinst
> > was trying to start hal again and it then hangs.  I can see 2
> > hald's running.  Running /etc/init.d/hal stop olny stopped the
> > second one.  I had to manually kill the first one.  After this
> > it installed properly.
> 
> Can you reproduce the problem reliably?

I have no idea.  I'm not in the habbit of doing dist-upgrades. :)

> The old hald should be stopped in the prerm script, could you check if this is
> the case.
> Is there maybe a hal trigger activated in the middle of the upgrade which
> restarts hal? (a screen log would be helpful) Otherwise I don't see how hald
> could be running.

The dpkg.log file says:
2009-12-17 09:02:07 upgrade libhal1 0.5.11-8 0.5.13-6
2009-12-17 09:02:07 status half-configured libhal1 0.5.11-8
2009-12-17 09:02:07 status unpacked libhal1 0.5.11-8
2009-12-17 09:02:07 status half-installed libhal1 0.5.11-8
2009-12-17 09:02:07 status half-installed libhal1 0.5.11-8
2009-12-17 09:02:07 status unpacked libhal1 0.5.13-6
2009-12-17 09:02:07 status unpacked libhal1 0.5.13-6
2009-12-17 09:02:07 upgrade libhal-storage1 0.5.11-8 0.5.13-6
2009-12-17 09:02:07 status half-configured libhal-storage1 0.5.11-8
2009-12-17 09:02:07 status unpacked libhal-storage1 0.5.11-8
2009-12-17 09:02:07 status half-installed libhal-storage1 0.5.11-8
2009-12-17 09:02:07 status half-installed libhal-storage1 0.5.11-8
2009-12-17 09:02:07 status unpacked libhal-storage1 0.5.13-6
2009-12-17 09:02:07 status unpacked libhal-storage1 0.5.13-6
[...]
2009-12-17 09:02:30 status half-installed xserver-xorg-input-wacom 0.8.0.2-2
2009-12-17 09:02:31 status triggers-pending hal 0.5.11-8
2009-12-17 09:02:31 status half-installed xserver-xorg-input-wacom 0.8.0.2-2
[...]
2009-12-17 09:02:49 trigproc hal 0.5.11-8 0.5.11-8
2009-12-17 09:02:49 status half-configured hal 0.5.11-8
2009-12-17 09:02:52 status installed hal 0.5.11-8
[...]
2009-12-17 09:03:47 upgrade udev 0.125-7+lenny3 149-1
2009-12-17 09:03:47 status half-configured udev 0.125-7+lenny3
2009-12-17 09:03:48 status unpacked udev 0.125-7+lenny3
2009-12-17 09:03:48 status half-installed udev 0.125-7+lenny3
2009-12-17 09:03:51 status triggers-pending doc-base 0.8.20
2009-12-17 09:03:51 status half-installed udev 0.125-7+lenny3
2009-12-17 09:03:51 status triggers-pending man-db 2.5.2-4
2009-12-17 09:03:51 status half-installed udev 0.125-7+lenny3
2009-12-17 09:03:51 status half-installed udev 0.125-7+lenny3
2009-12-17 09:03:52 status unpacked udev 149-1
2009-12-17 09:03:52 status unpacked udev 149-1
2009-12-17 09:03:52 upgrade hal-info 20080508+git20080601-1 20091130-1
2009-12-17 09:03:52 status half-configured hal-info 20080508+git20080601-1
2009-12-17 09:03:52 status unpacked hal-info 20080508+git20080601-1
2009-12-17 09:03:52 status half-installed hal-info 20080508+git20080601-1
2009-12-17 09:03:52 status triggers-pending hal 0.5.11-8
2009-12-17 09:03:52 status half-installed hal-info 20080508+git20080601-1
2009-12-17 09:03:52 status half-installed hal-info 20080508+git20080601-1
2009-12-17 09:03:52 status unpacked hal-info 20091130-1
2009-12-17 09:03:52 status unpacked hal-info 20091130-1
2009-12-17 09:03:52 upgrade hal 0.5.11-8 0.5.13-6
2009-12-17 09:03:52 status half-configured hal 0.5.11-8
2009-12-17 09:03:52 status unpacked hal 0.5.11-8
2009-12-17 09:03:52 status half-installed hal 0.5.11-8
2009-12-17 09:03:53 status half-installed hal 0.5.11-8
2009-12-17 09:03:53 status half-installed hal 0.5.11-8
2009-12-17 09:03:53 status unpacked hal 0.5.13-6
2009-12-17 09:03:53 status unpacked hal 0.5.13-6
[...]
2009-12-17 19:03:16 configure libhal1 0.5.13-6 0.5.13-6
2009-12-17 19:03:16 status unpacked libhal1 0.5.13-6
2009-12-17 19:03:16 status half-configured libhal1 0.5.13-6
2009-12-17 19:03:16 status installed libhal1 0.5.13-6
2009-12-17 19:03:16 configure libhal-storage1 0.5.13-6 0.5.13-6
2009-12-17 19:03:16 status unpacked libhal-storage1 0.5.13-6
2009-12-17 19:03:16 status half-configured libhal-storage1 0.5.13-6
2009-12-17 19:03:16 status installed libhal-storage1 0.5.13-6
[...]
2009-12-17 19:08:44 configure hal-info 20091130-1 20091130-1
2009-12-17 19:08:44 status unpacked hal-info 20091130-1
2009-12-17 19:08:44 status half-configured hal-info 20091130-1
2009-12-17 19:08:44 status installed hal-info 20091130-1
2009-12-17 19:08:44 configure hal 0.5.13-6 0.5.13-6
2009-12-17 19:08:44 status unpacked hal 0.5.13-6
2009-12-17 19:08:44 status unpacked hal 0.5.13-6
2009-12-17 19:08:44 status unpacked hal 0.5.13-6
2009-12-17 19:08:44 status unpacked hal 0.5.13-6
2009-12-17 19:08:44 status unpacked hal 0.5.13-6
2009-12-17 19:08:44 status half-configured hal 0.5.13-6

So it look like it was triggered before being unpacked,
but nothing happened with that.

But from apt's log, it seems that dbus restarted hal:
Setting up dbus (1.2.16-2) ...
Installing new version of config file /etc/init.d/dbus ...
Installing new version of config file /etc/dbus-1/system.conf ...
Installing new version of config file /etc/dbus-1/session.conf ...
Stopping Hardware abstraction layer: hald.
Stopping system message bus: dbus.
Starting system message bus: dbus.
Starting Hardware abstraction layer: hald.
Setting up libudev0 (149-1) ...
Setting up udev (149-1) ...

Which in dpkg.log seems to be:
2009-12-17 18:59:36 status installed adduser 3.111
2009-12-17 18:59:36 configure dbus 1.2.16-2 1.2.16-2
2009-12-17 18:59:36 status unpacked dbus 1.2.16-2
2009-12-17 18:59:36 status unpacked dbus 1.2.16-2
2009-12-17 18:59:36 status unpacked dbus 1.2.16-2
2009-12-17 18:59:36 status unpacked dbus 1.2.16-2
2009-12-17 18:59:36 status unpacked dbus 1.2.16-2
2009-12-17 18:59:36 status half-configured dbus 1.2.16-2
2009-12-17 18:59:45 status installed dbus 1.2.16-2
2009-12-17 18:59:45 configure libudev0 149-1 149-1

So that was between unpacking everything and where things got
configured.

If my memory is any good, ps indicated that hald was started
around 18:59, so that does seem to match.

> Another interesting point is, why starting a second instance of hald is actually
> successful.
> I'm using start-stop-daemon --start --oknodo --pidfile in the init script which
> should prevent a second instance from being started.

My guess would be that the pid file contains something wrong or
wasn't present.


Kurt