Crash after initializing GLX on -nvidia

Yep, me too after I upgraded today. Seems to stop at initialising 'winbind daemon'

having the same problem, however after a long wait ( 5 min or more ) login screen appears and I can login without a problem it seems.

I have a similar problem. I upgraded 2 days ago, after that I can't boot. I tried different kernels and recover mode, but they all have the same problem. My log files are not updated with the failed boots. The last msg I see is 'clocksource tsc unstable'. I am able to press ctrl-alt-del to reboot. I let my system set for a couple hr to see if the login would appear, but it didn't.

How can I recover from this without formatting or reinstalling?

Setting to critical, as it appears multiple users are experiencing boot failures.

flyapen: your issue is probably bug #431812 .

The original report by "Christopher" seems to be a duplicate of #429003 -- at least hist Xorg.log shows xorg froze when initializing the nvidia GLX module.

I've seen this a couple times, and am attaching my xorg log from the last time it happened - it also looks like the nvidia driver is failing to initialise - could this be some weird usplash / xsplash conflict where usplash still has the framebuffer but x (via the nvidia driver) is now trying to grab it??

Alex, usplash is disabled at boot, so you should not be seeing it all. With that said, I suggest opening a new bug on this against xorg, by running 'ubuntu-bug xorg'. I have an Nvidia graphics chipset on my AMD64 box and do not experience this issue, but I use an analog monitor...and I noticed you are using a digital connection. Could be the difference.

Hello. I to start ubuntu without loging, but sometimes it is hung on the screen of take-off (attached picture

I forgot to comment that I am affected by the bug "powernow-k8" by a motherboard MSI AM2 Platform

Thans by your attention

We recently patched an issue with xsplash that was causing it to hang on all UNR installations. If this happens to you, please try a Ctrl-Alt-F1 to a terminal screen. Assuming you can do this and login, you should be able to kill xsplash, and then run:
1) sudo apt-get update
2) sudo apt-get dist-upgrade

and then retry the boot.

The xsplash fix was connected to hang after login; it doesn't sound like this matches the original bug report here.

hmm...good point

@Christopher: As the original reporter, could you download Karmic Beta (when it releases today) and boot into the LiveCD and let us know if you observe any boot issues? If not, we can then provide instructions on how to update your system to the latest packages so that you can boot.

I must say that this sounds like X failing to start, rather than like a boot sequence problem?

Xers, could you please investigate asap and add a task to the correct package if needed. Thanks.

I used the Karmic Beta livecd and the screen flickers a little bit, then stops outputting video and the system seems to hang. I have an NV gfx card, amd64 livecd. I dl'd the ISO and used the USB creator and booted from a flash drive.

For what it is worth, I had a similar problem after an 'apt-get upgrade' a few days ago. I booted, saw the B/W ubuntu logo, and then the screen changed to a text mode, showed some services starting, and then started flickering incessantly. I was unable to change to a different VT and log in (it was missing key inputs - presumably because it was switching back to VT7 for X).

Unlike the others here, my Xlog.0 file indicated that the (restricted) Nvidia drivers weren't able to load. I was able to boot using an older kernel selection from Grub. I was also able to boot using the Beta .iso image on an USB stick.

It seems that for some reason, the nvidia drivers weren't rebuilt when the kernel was upgraded during the 'apt-get upgrade' I ran yesterday to move from Alpha 6 to Beta 1. A 'sudo dpkg-reconfigure nvidia-185-kernel-source' ended up fixing the problem for me.

I only mention it as it seems that many of the problem systems are using Nvidia chipsets. At this point, I'm assuming it was a one time glitch and don't plan to file a bug report. Please let me know if I should do otherwise.

I had the same problem after upgrading to karmic beta.
Booting with the older kernel worked perfectly
Reinstalling the nvidia propriety drivers sovled my prboblem

So I should uninstall nvidia stuff using safe mode? :o

I'm getting the same problem with a Lenovo S12 with Intel 945GME graphics chipset. How can I help debug this? When the screen is blank, I can't do a "control-alt-F1" to get to a different shell to diagnose the problem.

passing "apci=off" fixes the problem for me... (daily Karmic build from Oct. 5 on lenovo s12 intel atom)... will diagnose it more....

I updated my system, backed up my home folder and ran the following to back up my installations:

dpkg --get-selections > ~/my-packages

I saved it to my home folder and ran the following after a clean install of the beta.

sudo dpkg --set selections < ~my-packages && sudo apt-get dselect-upgrade

This will download and install all your applications, except those manually installed, although they may be in the repository.

It worked for me, and the new install is working as it should.

Since people haven't been including their Xorg.0.log's it's impossible to say whether people actually are seeing the same bug or not. "Hangs before login" is kind of a generic symptom that could be caused by any number of different underlying issues.

Focusing on the original reporter's issue, the Xorg.0.log shows that this is an X crash involving the -nvidia driver. Unfortunately the log does not show the actual backtrace, just "Backtrace:". Christopher, you'll need to collect a manual backtrace - see http://wiki.ubuntu.com/X/Backtracing for details. I'm not sure that apport will be able to collect it in this case, you may need to get it with gdb.

[Dropping targeting this to Karmic. -nvidia is a proprietary binary blob, so even if we knew exactly what the problem is we couldn't fix it since we don't have access to the source code. We will have to wait for Nvidia to put out a fixed version, which is not something we can expect to happen in time for Karmic.]

Btw, aside from a backtrace, there's another technique that could be used to trace this bug down. In the original report it suggests that the bug appeared only after doing some updates, implying that things worked properly prior to that. So check your /var/log/dpkg.log to see what packages changed, and try downgrading suspicious looking packages until you identify the one that caused it.

also, I see no problem when I pass "nosplash" on the kernel command line (set using grub)...

i have the same problem like Ned's description but the dpkg-reconfigure did not help i tried the nv 185 and 190beta same results it first appears at an update in the 2.6.31.10 kernel 2-3weeks ago

ubuntu64bit
amd64 old single core 3500+
chipset nv4 or 5
nv7800gt

Xorg.log only says cant load nv kernel module see kernel.log but in there i found no hint

sorry i dont have the upgrade in my dpkg.log but it was in the livetime of the 2.6.31.10 kernel

Hi,

May someone have a look and confirm if the bug I report match this one ?
Bug #454938 I don't have the expertise the understand this kind of crash.

Hello I am attempting to install 9.10 final to this hardware MSI NX7600GS-TD256 GeForce 7600GS 256MB 128-bit GDDR2 AGP 4X/8X Video Card All I get is out of range. I have attempted many things but I still can not get Karmic to boot successfully.

I had the same symptoms as Ned Patterson. The problem ceased after removing the /etc/X11/xorg.conf file, creating a new one (by executing nvidia-xconfig) and installing 185 kernel module (aptitude install nvidia-185-kernel-source).

I think karmic should do something automatically (or at least switch to nonrestricted drivers in problematic case), because doing it manually is too complicated for a beginning user -- he would rather give up and switch to windows... :(

M.

I have the same problem with nvidia driver 96 after update today. Here is my Xorg.log

I have libc6 2.10.1-0ubuntu15

Had to start in safe mode and remove xorg.conf to start with dri disabled adn nv driver, to use the pc again. But i'd like to have the dri back...

I have the same problem: with ubuntu 9.10 and nvidia 185 driver after grub the screen became black ("no vga signal"). after that i'm unable to switch to ctrl+alt+F1 or reboot with ctrl+alt+canc and I must restart computer.

My desktop is OK now with the 185 driver. I need to turn of and on the Acer 1715 monitor a few times to bring it back to life, but that's been an issue since intrepid.

Chris, okay good to hear it is working properly now. It would be nice to investigate this further but without a full backtrace we can't investigate further, so I guess I'll close this out at this time.

If anyone does get a full backtrace, please file a new bug (use ubuntu-bug xorg) and we'll check into it.

Hi Bryce,

I have what appears to be the same problem after an upgrade. I reported Bug #478727. At the current stage of managing a boot after 30 min with various tries I cannot be sure it is nvidia xorg. I have nvidia 185 drivers.

Thanks
Karim