[gm45] computer becomes unresponsive a few minutes after booting (karmic)

[gbh]

I'm on Karmic Koala. This problem manifested with numerous 2.6.31 kernels (64-bit).

I boot up, start a browser, browse around, then the screen freezes. I can only move the mouse cursor, and it's quite jerky. I cannot get to a text console. I am able to ssh in. After trying many processes, I've found that killing X brings me back to the log-in screen, and the machine is responsive again. I log back in, and in a few minutes the same thing happens.

With some experimenting, I found it to be proximately induced by a browser tab with a large (though not gargantuan) image.

I was not able to get any debugging information. Nothing showed up in /var/log/messages. I tested the memory and the hard drive. I tried a live CD without problems.

I had a few hours earlier a routine update from synaptic, consisting of the following:

2010-05-27 06:41:52 upgrade libc-dev-bin 2.10.1-0ubuntu16 2.10.1-0ubuntu17
2010-05-27 06:41:54 upgrade libc6-dev 2.10.1-0ubuntu16 2.10.1-0ubuntu17
2010-05-27 06:41:56 upgrade libc6-dev-i386 2.10.1-0ubuntu16 2.10.1-0ubuntu17
2010-05-27 06:41:57 upgrade libc6-i386 2.10.1-0ubuntu16 2.10.1-0ubuntu17
2010-05-27 06:41:58 upgrade libc-bin 2.10.1-0ubuntu16 2.10.1-0ubuntu17
2010-05-27 06:42:04 upgrade libc6 2.10.1-0ubuntu16 2.10.1-0ubuntu17
2010-05-27 06:42:18 upgrade chromium-codecs-ffmpeg-extra 0.5+svn20100430r46138+46111+47191-0ubuntu1~ucd1~karmic 0.5+svn20100522r47988+47943+48268-0ubuntu1~ucd1~karmic
2010-05-27 06:42:19 upgrade chromium-browser-inspector 6.0.404.0~svn20100514r47231-0ubuntu1~ucd1~karmic 6.0.417.0~svn20100527r48348-0ubuntu1~ucd1~karmic
2010-05-27 06:42:20 upgrade chromium-browser 6.0.404.0~svn20100514r47231-0ubuntu1~ucd1~karmic 6.0.417.0~svn20100527r48348-0ubuntu1~ucd1~karmic
2010-05-27 06:42:29 upgrade google-chrome-beta 5.0.375.38-r46659 5.0.375.55-r47796
2010-05-27 06:42:36 upgrade libkrb5-dev 1.7dfsg~beta3-1ubuntu0.5 1.7dfsg~beta3-1ubuntu0.6
2010-05-27 06:42:36 upgrade libgssapi-krb5-2 1.7dfsg~beta3-1ubuntu0.5 1.7dfsg~beta3-1ubuntu0.6
2010-05-27 06:42:37 upgrade libkrb5-3 1.7dfsg~beta3-1ubuntu0.5 1.7dfsg~beta3-1ubuntu0.6
2010-05-27 06:42:38 upgrade libkrb5support0 1.7dfsg~beta3-1ubuntu0.5 1.7dfsg~beta3-1ubuntu0.6
2010-05-27 06:42:39 upgrade libk5crypto3 1.7dfsg~beta3-1ubuntu0.5 1.7dfsg~beta3-1ubuntu0.6
2010-05-27 06:42:39 upgrade libgssrpc4 1.7dfsg~beta3-1ubuntu0.5 1.7dfsg~beta3-1ubuntu0.6
2010-05-27 06:42:40 upgrade libkdb5-4 1.7dfsg~beta3-1ubuntu0.5 1.7dfsg~beta3-1ubuntu0.6
2010-05-27 06:42:40 upgrade libkadm5srv6 1.7dfsg~beta3-1ubuntu0.5 1.7dfsg~beta3-1ubuntu0.6
2010-05-27 06:42:41 upgrade libkadm5clnt6 1.7dfsg~beta3-1ubuntu0.5 1.7dfsg~beta3-1ubuntu0.6
2010-05-27 06:42:42 upgrade chromium-codecs-ffmpeg-nonfree 0.5+svn20100430r46138+46111+47191-0ubuntu1~ucd1~karmic 0.5+svn20100522r47988+47943+48268-0ubuntu1~ucd1~karmic
2010-05-27 06:42:42 upgrade gdm 2.28.1-0ubuntu2.1 2.28.1-0ubuntu2.2
2010-05-27 06:42:53 upgrade libpq-dev 8.4.3-0ubuntu9.10.1 8.4.4-0ubuntu9.10
2010-05-27 06:42:54 upgrade libpq5 8.4.3-0ubuntu9.10.1 8.4.4-0ubuntu9.10
2010-05-27 06:42:55 upgrade xserver-common 2:1.6.4-2ubuntu4.2 2:1.6.4-2ubuntu4.3
2010-05-27 06:42:56 upgrade xserver-xorg-core 2:1.6.4-2ubuntu4.2 2:1.6.4-2ubuntu4.3
2010-05-27 06:42:57 upgrade xvfb 2:1.6.4-2ubuntu4.2 2:1.6.4-2ubuntu4.3

It had been weeks since I'd done an update before that.

Since it appeared to be X-related (and top usually showed Xorg eating CPU), I tried reverting the X packages back (well, back to a version I could readily get). This did not help.

I started reverting seemingly irrelevant packages, including the linux-headers and the libc packages. I went back to a much earlier kernel version. The problem appears to have gone away.

My best guess (based on likelihood) is that it's a problem with the libc packages, or at least their interaction with X. If so, it's quite serious, as it rendered the machine basically unusable. I hope others can avoid this nightmare.

(Unrelatedly, Launchpad wouldn't let me submit this because I set my browser to not send Referer headers. No problem, I just turn them back on, right? No, I hit the back button and the whole bug report was gone. Thanks, Launchpad! So I re-enter it. Next, timeout errors!)

Update: Never mind about it being fixed, it happened again, it just took longer.

[Bryce Harrington (bryce)]

Hi gbh,

Please run the command 'apport-collect BUGNUMBER', which will attach several files we need for debugging.

[This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

[gbh (galen-math)]

Architecture: amd64
DistroRelease: Ubuntu 9.10
InstallationMedia: Ubuntu 9.10 "Karmic Koala" - Release amd64 (20091027)
MachineType: Acer Aspire 1410
Package: xorg 1:7.4+3ubuntu10
PackageArchitecture: amd64
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.31-21-generic root=UUID=430af2b1-d2e9-4351-bf64-2d40761efe6c ro quiet splash
ProcEnviron:
 SHELL=/bin/bash
 PATH=(custom, user)
 LANG=en_US.UTF-8
ProcVersionSignature: Ubuntu 2.6.31-21.59-generic
RelatedPackageVersions:
 xserver-xorg 1:7.4+3ubuntu10
 libgl1-mesa-glx 7.6.0-1ubuntu4
 libdrm2 2.4.14-1ubuntu1
 xserver-xorg-video-intel 2:2.9.0-1ubuntu2.1
 xserver-xorg-video-ati 1:6.12.99+git20090929.7968e1fb-0ubuntu1
Uname: Linux 2.6.31-21-generic x86_64
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare
XorgConf: Error: [Errno 2] No such file or directory: '/etc/X11/xorg.conf'
dmi.bios.date: 08/14/2009
dmi.bios.vendor: INSYDE
dmi.bios.version: v0.3115
dmi.board.asset.tag: Base Board Asset Tag
dmi.board.name: Base Board Product Name
dmi.board.vendor: Acer
dmi.board.version: Base Board Version
dmi.chassis.type: 1
dmi.chassis.vendor: Chassis Manufacturer
dmi.chassis.version: Chassis Version
dmi.modalias: dmi:bvnINSYDE:bvrv0.3115:bd08/14/2009:svnAcer:pnAspire1410:pvrv0.3115:rvnAcer:rnBaseBoardProductName:rvrBaseBoardVersion:cvnChassisManufacturer:ct1:cvrChassisVersion:
dmi.product.name: Aspire 1410
dmi.product.version: v0.3115
dmi.sys.vendor: Acer
fglrx: Not loaded
system:
 distro: Ubuntu
 architecture: x86_64kernel: 2.6.31-21-generic

[gbh (galen-math)]

More info: I have tried thrashing this computer every way I can think of: massive swapping, multiple high-CPU processes, large network load, etc. None of these cause a crash. Its "berserk button" is this one page:

http://www.snpedia.com/index.php/Haplogroup_H_(mtDNA)

Opening it in Firefox usually induces an immediate crash with the symptoms described above. By removing that tab and not viewing it again, the problem has not returned for several weeks. To be safe, I'm not updating any packages for now.

[Stenten (stenten)]

Do you still experience this bug if you don't suspend/hibernate, i.e. immediately after a fresh boot?

[gbh (galen-math)]

Yes immediately.

[Stenten (stenten)]

Could you please reproduce this issue on a fresh boot and then either switch to a TTY or SSH in and get a copy of /var/log/Xorg.0.log, /var/log/dmesg, and the output of dmesg?

Also please test with the xorg-edgers PPA [1]. Make sure you install 'ppa-purge' before upgrading so you can revert any changes if needed.

Also try the current mainline kernel [2]. Installation instructions are here [3].

[1]: https://launchpad.net/~xorg-edgers/+archive/ppa
[2]: http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.35-rc3-maverick/
[3]: https://wiki.ubuntu.com/KernelTeam/MainlineBuilds

[Launchpad Janitor (janitor)]

[Expired for xserver-xorg-video-intel (Ubuntu) because there has been no activity for 60 days.]