X crashes while upgrading from Feisty to Gutsy

[David MENTRÉ]

Hello,

While upgrading from Feisty to Gutsy, the X server crashed near the end of the upgrade (36 packages remaining to install on about 1000 packages). After trying to reconnect, the upgrade system told me to reboot the machine, which I did. The reboot went fine and I could log in the system. But after trying to update the system with the update manager (System -> Administration -> "Gestionnaire de mise à jour" in French) the update manager tells me that the repository is corrupted. Doing a "sudo dpkg --configure -a" and then a reboot seems to fix the issue.

I'm using a regular Ubuntu install, on a desktop system.

Sincerely yours,
david

[jtholmes (jtholmes)]

changed to x11-common

[jtholmes (jtholmes)]

thank you for your report

you said dpkg --configure -a fixed the issue

is there still a problem or can we close this report

[David MENTRÉ (dmentre)]

Hello,

jtholmes <email address hidden> writes:

> you said dpkg --configure -a fixed the issue
>
> is there still a problem or can we close this report

There is no longer a problem on my machine.

However, I do think the issue I reported might show a more problematic
underlying issue and that you should *not* close the report.

The upgrade should not have crashed the X server at first. After reboot,
the system was in an unusable state: the automatic update system refused
to finish the upgrade and aborted with an English error message in
cryptic form. For an non-English speaking person and less technically
knowledgeable than me (well, I mean a standard user), that would have
been a show stopper.

Yours,
david

[Timo Aaltonen (tjaalton)]

David: please attach your /var/log/dpkg.log*.gz that has the log entries of the time you did the upgrade. The logs are rotated monthly.

[David MENTRÉ (dmentre)]

Timo: here it is. I did the upgrade on 2007-12-26.

[Timo Aaltonen (tjaalton)]

2007-12-26 14:12:39 status unpacked libgtkhtml3.8-15 1:3.13.5-1
2007-12-26 14:12:39 status half-configured libgtkhtml3.8-15 1:3.13.5-1
<crashed here>
2007-12-26 14:16:55 startup packages configure
2007-12-26 14:16:55 trigproc initramfs-tools 0.85eubuntu20 0.85eubuntu20
2007-12-26 14:16:55 status half-configured initramfs-tools 0.85eubuntu20

The log wasn't that helpful.. Unless you have ideas I can only close the bug :) Possibly you were just unlucky that the crash happened during update. It could be a hardware problem that makes the system crash under load (the update progressed for about an hour, which means intense disk access and cpu usage).

[David MENTRÉ (dmentre)]

Hello Timo,

Timo Aaltonen <email address hidden> writes:

> The log wasn't that helpful.. Unless you have ideas I can only close the
> bug :)

Could the screen saver might have played a role? I wasn't in front of
the computer during the whole update so the screen saver has started
several times.

Otherwise, I've gone through the log files and found this in
/var/log/apport.log.4.gz:

apport (pid 20697) Wed Dec 26 14:12:25 2007: called for pid 5498, signal 11
apport (pid 20697) Wed Dec 26 14:12:25 2007: Unhandled exception:
Traceback (most recent call last):
  File "/usr/share/apport/apport", line 275, in <module>
    info.add_proc_info(pid)
  File "/var/lib/python-support/python2.5/apport/report.py", line 368, in add_proc_info
    assert os.path.exists(self['ExecutablePath'])
  File "posixpath.py", line 171, in exists
    st = os.stat(path)
TypeError: stat() argument 1 must be (encoded string without NULL bytes), not str
pid: 20697, uid: 1000, gid: 1000, euid: 0, egid: 0
environment: {'CORE_GID': '1000', 'CORE_TIME': '1198674744', 'CORE_EXECUTABLE': 'gnome-panel', 'CORE_HOSTNAME': 'morgana', 'CORE_SIGNAL': '11', 'CORE_REAL_RLIM': '0', 'CORE_PID': '5498', 'CORE_UID': '1000'}
apport (pid 20734) Wed Dec 26 14:12:29 2007: called for pid 4830, signal 6
apport (pid 20734) Wed Dec 26 14:12:29 2007: Unhandled exception:
Traceback (most recent call last):
  File "/usr/share/apport/apport", line 275, in <module>
    info.add_proc_info(pid)
  File "/var/lib/python-support/python2.5/apport/report.py", line 368, in add_proc_info
    assert os.path.exists(self['ExecutablePath'])
  File "posixpath.py", line 171, in exists
    st = os.stat(path)
TypeError: stat() argument 1 must be (encoded string without NULL bytes), not str
pid: 20734, uid: 0, gid: 0, euid: 0, egid: 0
environment: {'CORE_GID': '0', 'CORE_TIME': '1198674748', 'CORE_EXECUTABLE': 'Xorg', 'CORE_HOSTNAME': 'morgana', 'CORE_SIGNAL': '6', 'CORE_REAL_RLIM': '0', 'CORE_PID': '4830', 'CORE_UID': '0'}

Could it be related, especially the second report?

> Possibly you were just unlucky that the crash happened during
> update. It could be a hardware problem that makes the system crash under
> load (the update progressed for about an hour, which means intense disk
> access and cpu usage).

Maybe. But I had this hardware for years and never had an issue with it.

Yours,
d.

[Timo Aaltonen (tjaalton)]

Oh yeah, screensaver could have done it. Maybe update-manager should ensure to not let the screensaver start.

[Andreas Moog (ampelbein)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. However, you are not using the most recent version of the Ubuntu release. Please upgrade to the most recent version and when you are still having this issue, feel free to file a new bug-report. Thanks in advance.