xserver-common config script hangs on upgrade

Bug #8732 reported by Luc Stroobant
6
Affects Status Importance Assigned to Milestone
Ubuntu
Invalid
Medium
Matt Zimmerman

Bug Description

Since the last upgrade for xfree86, my apt-get hangs. This is the output:

Extracting templates from packages: 100%
Preconfiguring packages ...
(no more reaction, even after af few hours)

When I "control-c" apt-get, I get:

Received signal. Aborting xserver-xfree86 package config script.

Which makes me think this is an xfree86 problem? (I can install or upgrade other
packes with apt-get install <name>)

I'm using a Via C3-2 cpu - M10000 system, specs here:
http://www.epiacenter.com/modules.php?name=Content&pa=showpage&pid=21

http://www.stroobant.be/~luc/dpkg.strace: http://www.stroobant.be/~luc/dpkg.strace

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

From which version to which verision of X are you trying to upgrade?
I have never see such a problem in all my tests.
Try also this:

export DEBUG_XFREE86_PACKAGE=yes
apt-get update
apt-get upgrade

There will be more output than usual. Please collect it and attach it to the bug.

Fabio

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

apt-get upgrade output:

Extracting templates from packages: 100%
Preconfiguring packages ...
xserver-common config note: Linux kernel > 2.5 detected; using 0 as default
   nice value
xserver-common config note: setting xserver-common/xwrapper/allowed_users
   from configuration file
xserver-common config note: setting xserver-common/xwrapper/nice_value from
   configuration file
xserver-common config note: command "db_input low
   xserver-common/xwrapper/allowed_users" exited with status 30
xserver-common config note: command "db_input low
   xserver-common/xwrapper/nice_value" exited with status 30

...
(hangs here)

# XFree86 -version

This is a pre-release version of XFree86, and is not supported in any
way. Bugs may be reported to <email address hidden> and patches submitted
to <email address hidden>. Before reporting bugs in pre-release versions,
please check the latest version in the XFree86 CVS repository
(http://www.XFree86.Org/cvs).

XFree86 Version 4.3.0.1 (Ubuntu 4.3.0.dfsg.1-6ubuntu21 20040922153708
<email address hidden>)
Release Date: 15 August 2003
X Protocol Version 11, Revision 0, Release 6.6
Build Operating System: Linux 2.4.27-es i686 [ELF]
Build Date: 22 September 2004
        Before reporting problems, check http://www.XFree86.Org/
        to make sure that you have the latest version.
Module Loader present
OS Kernel: Linux version 2.6.8.1-2-386 (buildd@mcmurdo) (gcc version 3.3.4
(Debian 1:3.3.4-9ubuntu5)) #1 Sat Sep 18 09:49:53 BST 2004

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

Can you please do also the following steps? I really can figure out what is wrong.

export DEBUG_XFREE86_PACKAGE=yes

dpkg-reconfigure xserver-common

and see if it still hangs or loop somewhere.

Thanks
Fabio

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

Reassing the bug to proper component.

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

dpkg-reconfigure works fine. I can set the users allowed to start X and the nice
value.

But a new apt-get update / apt-get upgrade gives the same problem:

Extracting templates from packages: 100%
Preconfiguring packages ...
xserver-common config note: Linux kernel > 2.5 detected; using 0 as default
   nice value
xserver-common config note: setting xserver-common/xwrapper/allowed_users
   from configuration file
xserver-common config note: setting xserver-common/xwrapper/nice_value from
   configuration file
xserver-common config note: command "db_input low
   xserver-common/xwrapper/allowed_users" exited with status 30
xserver-common config note: command "db_input low
   xserver-common/xwrapper/nice_value" exited with status 30

(PS: I can give you an account on the machine if that helps, just send me an
email...)

Revision history for this message
Daniel Stone (daniels) wrote :

I strongly believe this is a local problem; we have not experienced the problem
within Ubuntu before, and the same script is used in Debian with no ill effects.
 What filesystem are you using?

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

Here is my fstab:
(All system partitions have been erased during the ubuntu install. )

# <file system> <mount point> <type> <options> <dump> <pass>
proc /proc proc defaults 0 0
/dev/hda1 / ext3 defaults,errors=remount-ro 0 1
/dev/hda6 /home ext3 defaults 0 2
/dev/hda5 /tmp reiserfs defaults 0 2
/dev/hda7 /usr reiserfs defaults 0 2
/dev/hda8 /var reiserfs defaults 0 2
/dev/hda9 none swap sw 0 0
/dev/scd0 /media/cdrom0 udf,iso9660 ro,user,noauto 0 0

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

Hi,
  please download this package:

http://people.ubuntulinux.org/~fabbione/xserver-common_4.3.0.dfsg.1-6ubuntu23_i386.deb

and install it using:

dpkg -i xserver-common_4.3.0.dfsg.1-6ubuntu23_i386.deb 2>log

and attach the log file to this bug. It should tell us theoretically where is
hanging.
Once we can spot the hang point we will try using a similar script to trim down
where
the real issue is.

Fabio

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

Created an attachment (id=343)
Logfile xserver-common_4.3.0.dfsg.1-6ubuntu23_i386.deb

This logfile is the out from:
# dpkg -i xserver-common_4.3.0.dfsg.1-6ubuntu23_i386.deb 2> log

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

This is definetely not an X problem. the last script exit 0 and that is correct.
It's the last thing X has to do.
At that point either dpkg or apt or whatever package should keep going and it
doesn't.

I really have no idea what could be.

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

Mmm, that's not good. :-(
Updating my system is quite difficult this way, because X is always the first
package he takes an that's where it hangs...
Should I file a new bug for apt-get/dpkg or do you have an other suggestion?

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

Perhaps you can try to strace the process...

strace dpkg -i xserver-common_4.3.0.dfsg.1-6ubuntu23_i386.deb 2> log

and see if something later is hanging.

Fabio

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

I'm not a developper and I don't know how these packages depend on each other,
but are you sure the problem is in the xserver_common package? As you said, this
seems to install correct if I use dpkg -i:

nehemia:/var/cache/apt/archives# dpkg -i
xserver-common_4.3.0.dfsg.1-6ubuntu23_i386.deb
(Database inlezen ... 68945 bestanden en mappen geïnstalleerd.)
Voorbereiden om xserver-common 4.3.0.dfsg.1-6ubuntu23 te vervangen (met
xserver-common_4.3.0.dfsg.1-6ubuntu23_i386.deb) ...
(done)

To me, it seems to be the xserver_xfree86 which hangs:

nehemia:/var/cache/apt/archives# dpkg -i
xserver-xfree86_4.3.0.dfsg.1-6ubuntu23_i386.deb
(Database inlezen ... 68945 bestanden en mappen geïnstalleerd.)
Voorbereiden om xserver-xfree86 4.3.0.dfsg.1-6ubuntu21 te vervangen (met
xserver-xfree86_4.3.0.dfsg.1-6ubuntu23_i386.deb) ...
(hangs here)

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

AH!

ok please export DEBUG_XFREE86_PACKAGE=yes and try to upgrade.

Thanks
Fabio

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

nehemia:/var/cache/apt/archives# dpkg -i
xserver-xfree86_4.3.0.dfsg.1-6ubuntu23_i386.deb 2>log
(Database inlezen ... 68749 bestanden en mappen geïnstalleerd.)
Uitpakken van xserver-xfree86 (uit
xserver-xfree86_4.3.0.dfsg.1-6ubuntu23_i386.deb) ...

(nothing happens - 1 hour later I hit CTRL-C)

nehemia:/var/cache/apt/archives# cat log

Received signal. Aborting xserver-xfree86 package config script.

dpkg: fout bij afhandelen van xserver-xfree86_4.3.0.dfsg.1-6ubuntu23_i386.deb
(--install):
 subproces pre-installation script werd gedood door signaal (Onderbroken)
Fouten gevonden tijdens behandelen van:
 xserver-xfree86_4.3.0.dfsg.1-6ubuntu23_i386.deb

I'm afraid all data in the log is caused by cancelling the installation.
Short translation: error installing .... script killed by signal (interrupted)
... errors found installing: ...

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

Ok.. we should be close to find the problem.

Please grab this package:

http://people.ubuntulinux.org/~fabbione/xserver-xfree86_4.3.0.dfsg.1-6ubuntu23_i386.deb

and install with

dpkg -i xserver-xfree86_4.3.0.dfsg.1-6ubuntu23_i386.deb 2>log

and as usual please attach the full log here.

I might need more info later on according to what shows up there.

Thanks
Fabio

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

Created an attachment (id=355)
log xserver-xfree86 install

As requested, the log from:
dpkg -i xserver-xfree86_4.3.0.dfsg.1-6ubuntu23_i386.deb 2>log

Again: I had to cancel (CTRL-C) the installation after a while.

(If your next question is what kind of video-card I'm using:
0000:01:00.0 VGA compatible controller: VIA Technologies, Inc. VT8623 [Apollo
CLE266] integrated CastleRock graphics (rev 03))

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

Hi Luc,
       i am sorry but i really can't understand why it is failing.
I need you to try the last thing for me and i know it is not the correct way to
fix problems
but i don't think we have other options:

apt-get --purge remove xserver-xfree86
apt-get update
apt-get upgrade
apt-get install xserver-xfree86

after you have done these steps try again:

apt-get --reinstall install xserver-xfree86

and tell me if the script still hangs.

The only thing i am sure that this problem is related to something on your system,
otherwise tons of people would have reported the same issue.

Thanks
Fabio

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

Created an attachment (id=372)
lspci and cpuinfo

Fabio, the bad news is that I already tried that yesterday. It doesn't work and
still hangs during setup. :-(
(So my X is gone now, I hope the system doesn't crash before I get this is
fixed...)

As you say, my system is a bit special: it's a Via CPU and chipset. I'm quite
sure there are not much people using this a desktop. So it might be possible
that I'm the first Ubuntu user with this hardware. (but I've been running
Mandrake 10 and Debian stable/testing on it before)
Other things that might be interesting:
-during the first ubuntu setup, I had to specify my videocard from a list as
"via". It was not auto-detected.
-the via CPU does not support all recent Intel instructions, some packages that
are compiled for recent systems do not work on my system. (but that should not
be a problem, because Ubuntu is i386?)

I attached an lscpi -v and cat /proc/cpuinfo to be sure you have all info...

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

the point is that it hangs on an extremely simple foo=""!

which shell are you using? would it be possible for you to reinstall the system
or switch to another filesysyem other than reiserfs?

Fabio

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

Shell = just standard bash. I didn't change any settings for the shell as far as
I remember.
It's just a desktop, so a reinstall is possible. How easy/difficult is it to
change the filesystem without a reinstall? Can it be done from remote? (I'm
using SSH now -at work- and I'm afraid I don't have much time for a complete
reinstall next days...)

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

(In reply to comment #21)
> Shell = just standard bash. I didn't change any settings for the shell as far as
> I remember.

Ok you can just check where /bin/sh points to, but i guess it's bash.

> It's just a desktop, so a reinstall is possible. How easy/difficult is it to
> change the filesystem without a reinstall? Can it be done from remote? (I'm
> using SSH now -at work- and I'm afraid I don't have much time for a complete
> reinstall next days...)

Well you need to be able to copy all the data from the partition somewhere,
umount the partition, format it, change the /etc/fstab, remount it and copy
the data over again.

Theoretically it can be done remotly, until you don't lock yourself out ;)

Fabio

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

Fabio, since my xfree is in an "hanging" status, I've tried to uninstall it. Apt
suggest I have to install it before I can do an uninstall. I noticed this one in
the log when I do an strace from apt-get remove xserver-xfree86.

open("/usr/share/locale/nl_BE/LC_MESSAGES/apt.mo", O_RDONLY) = -1 ENOENT (No
such file or directory)
open("/usr/share/locale/nl/LC_MESSAGES/apt.mo", O_RDONLY) = 5

Is it possible that this issue is related to my locale settings? (Dutch -
Belgium nl_BE => azerty keyboard)

(Didn't try the format or reinstall yet, I don't have time for that now, sorry...)

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

Ok, this is weird. I rebooted my system, because I thought I had to do a
reinstall after all.
After the reboot, Ubuntu starts up and said something about x & xdm not started.
I log into the console as root, apt-get install xserver-common and everything
installs fine... Reboot once again and ubuntu starts up as expected. Everything
works fine now...

The only thing I did was a dpkg-reconfigure locales. This didn't seem to work
quite well, I got some errors on non-existing locale setting/files.

Locales with working config (nl_BE before that)
# locale
LANG=POSIX
LC_CTYPE="POSIX"
LC_NUMERIC="POSIX"
LC_TIME="POSIX"
LC_COLLATE="POSIX"
LC_MONETARY="POSIX"
LC_MESSAGES="POSIX"
LC_PAPER="POSIX"
LC_NAME="POSIX"
LC_ADDRESS="POSIX"
LC_TELEPHONE="POSIX"
LC_MEASUREMENT="POSIX"
LC_IDENTIFICATION="POSIX"
LC_ALL=

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

I believe that for one reason or another the system was in an inconsistent state
before the X update. Probably that might have cause the script to hang, but the
message
you straced is related to apt and not X. I wonder if somekind of disk corruption
might have take place. I am really guessing at this point in time.

Fabio

Revision history for this message
Esko Kauppinen (e-kauppinen) wrote :

Update: Got UBUNTU running by typing: "sudo xf86config" and choosing RADEON as
the graphics driver. Saved the file as /etc/X11/XF86Config-4.

Revision history for this message
Daniel Stone (daniels) wrote :

Uhm, looks like this is apt's bad, no?

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

Indeed, this seems to be an apt problem.
FYI, I had the same problem again with the recent X update (last week or something).
After a reboot, the update went fine...

Revision history for this message
Matt Zimmerman (mdz) wrote :

The problem was demonstrated with dpkg, so it is not an apt issue. It's likely
either a debconf or dpkg problem. We'll need a complete strace to find out
what's going wrong.

sudo strace -f -o /tmp/dpkg.strace dpkg -i <path to .deb which reproduces the
problem>

and attach /tmp/dpkg.strace

Revision history for this message
Matt Zimmerman (mdz) wrote :

This bug has been in need of more information for quite some time now, so I'm
closing it. Please reopen if anyone can provide the requested details.

Revision history for this message
Luc Stroobant (ubuntu-stroobant) wrote :

Matt, I was not able to provide this strace few months ago. Sorry for that, but
after a reboot it finally got installed and it was rather difficult to
reconstruct the problem.
Now I'm upgrading to hoary and the problem is back. Check the strace log on the
attached link. ( http://www.stroobant.be/~luc/dpkg.strace )

(I hope I don't have to reboot my system, I guess the install will work again
after a reboot, as it did last time.)

Revision history for this message
Matt Zimmerman (mdz) wrote :

Created an attachment (id=1317)
strace provided by submitter

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.