well yesterday i installed packages check and checkpolicy and havent had a freeze ever since, still unstable though
edit: nevermind, doenst work
Last edited by starhalo; June 3rd, 2009 at 09:49 PM.
There are a lot of Twitter users experiencing this problem (http://bit.ly/111eNn). For some of them, this is their first experience with Ubuntu and Linux. It's unfortunate that this bug is chipping away at the reputation of Ubuntu, even among those of us who have been using it for years. It's even worse that Ubuntu's shortcomings are reflecting on Linux as a whole.
With that said, I upgraded to the 2.6.29-02062903-generic kernel and haven't had any freezes since. It's too bad whatever fixes in that kernel haven't been pushed to Jaunty.
Last edited by utkjamie; June 4th, 2009 at 04:07 PM.
It's definitely something to investigate. Due the freezes i switched from 9.04 to 8.10 (very stable, on the same PC) and now i'm back to Jaunty... freezes are back too !
I noticed something : a "slow freeze" occurred while running the system monitor : one of the 2 CPU curves dropped down and sticked to the floor, and within 20 seconds everything was locked.
I tried to reproduce this situation, but without any success, as freezes are mostly instant ones. But i got it this morning for the second time.
As i said in my first post, nothing special seems to be corelated to the bug. It freeze with or without Wifi, nvidia driver etc.
The only thing that match nearly each freeze is a "clocksource tsc unstalble" :
Some weeks ago I tried alternate clock sources and NOTSC otions, but still got freezes.Code:kern.log : Jun 5 13:34:12 lapbuntu kernel: [ 86.000075] Clocksource tsc unstable (delta = -64193242 ns) messages : Jun 5 13:34:12 lapbuntu kernel: [ 86.000075] Clocksource tsc unstable (delta = -64193242 ns) syslog : Jun 5 13:34:12 lapbuntu kernel: [ 86.000075] Clocksource tsc unstable (delta = -64193242 ns) daemon.log : Jun 5 13:33:20 lapbuntu ntpdate[3685]: adjust time server 91.189.94.4 offset 0.470649 sec
Maybe we should call an exorcist ? Or a ghost buster ?
Last edited by Ormente; June 5th, 2009 at 01:17 PM.
Actually I thought the common problem is with Firefox/web use, possibly WiFi or Flash. And the symptoms generally seem in line with a graphics crash.
So tying everything together, it must be somewhere in the code that handles how dual-core processors transfer information between a WiFi input and the graphics drivers. Certain drivers seem to fare better than others too - in my experience, the open source ATi drivers are much more stable than the proprietary nVidia ones (no crashes at all under the former).
Either way, it's certainly reassuring to find this thread and realise it's happening to a lot of people... :-p
Big bump...
I've tried isolating this to alsa, pulseaudio, xserver-xorg-video-intel, X, firefox-3.xx, adobe flashplugin, compiz, pidgin, and NOTHING appears to solve this for me
Running a latitude D830 w intel graphics
Last edited by ozhoo; June 5th, 2009 at 09:22 PM.
I can't believe it's taken me two weeks of going demented over crashes and freezes to find this thread. I had no problems under 8.04, but since I upgraded to 9.04, I've had nothing but trouble. I have an Nvidia GeForce 7900 GS, and I originally thought the Nvidia drivers were causing the problem, because removing them completely seemed to fix it, but I've just had yet another crash, and am now unable to boot Ubuntu, even in recovery mode. I also have a Pentium Duo processor (3.4GHz).
Celeron 540M, Xubuntu 9.04
Hard Freeze today using Firefox while viewing flash heavy site. System completely unresponsive.
Only cure was ON/OFF
Follow up
My random freezing has stopped after updating to the latest kernel, currently @http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.30-rc8/
I don't have more than a few hours testing on my system, but one operation that would guarantee a freeze on my system with 2.6.28-11 does not freeze the system now. This is when I would try to decrypt a file with seahorse via nautilus.
I'll update again if I see the freezes continue, but users may want to try this fix. I would not uninstall your old kernels so you may go back to them if problems occur running with the new one.
I went through my logs as well. I didn't see this correlation that you pointed out with the "Clocksrouce tsc unstable" but now I'm watching for it. What I'm seeing is "hpet increasing min_delta_ns" entries is /var/log/kern.log and /var/log/messages:
CE: hpet increasing min_delta_ns to 15000 nsec
CE: hpet increasing min_delta_ns to 22500 nsec
CE: hpet increasing min_delta_ns to 33750 nsec
CE: hpet increasing min_delta_ns to 50624 nsec
After the last one it froze. So I'm grepping for both of these strings in kern.log and messages (also syslog, but nothing is there) to see if I can isolate that as a factor during a freeze.
This is either something with X or a kernel issue. I don't see anything in my X logs to indicate a problem though. My goal is to at least isolate that. Then take it from there. If it's a kernel issue, I'll try and roll back to an earlier version. If it's an X issue I might try going to an earlier version too.
Bookmarks