[gentoo-dev] System hangup w/ X taking 100% cpu

[gentoo-dev] System hangup w/ X taking 100% cpu

[Dan Armak]

Hi all,

On my new r12 system I've had a hangup in X11. No apparent cause. The system 
stopped accepting any input and the screen didn't even complete a redraw. The 
mouse went on moving as it was in h/w, but clicks were no good. Magic sysrq 
wasn't responding either.

I logged in via ssh, and top showed X was taking 100% cpu. I tried killall X, 
no effect. I tried kill -9 <X pid>, the remote login hung up and i couldn't 
ssh to it again. I had to do a hard reboot.

Then I expirienced the same effect twice on an old rc5 system which underwent 
absolutely no updates or config changes for a few months now. Only this was 
standalone not on a net, so I couldn't login via ssh.

Both rc6 lockups were when running kmail, both rc5 lockups while running 
ksirc.

Trying to send this message, I had another lockup on the rc6 system, hope 
this reaches you.

I'm absolutely mystified. Ideas? Anyone seen this kind of thing before?

-- 

Dan Armak
Gentoo Linux Developer, Desktop Team
Matan, Israel

Re: [gentoo-dev] System hangup w/ X taking 100% cpu

[djamil essaissi]

this is abviously a hardware problem, if you want help, we need to get
all the specs on your machines (the sick ones) : CPU, RAM MB ...
intel/amd speed NIC VideoCard.... etc

salam.


On Thu, 2001-12-13 at 19:19, Dan Armak wrote:
> > Trying to send this message, I had another lockup on the rc6 system, hope
> > this reaches you.
> I've had a 3rd lockup right after sending this off and  am now working on the 
> rc5 machine again. The rc6 one worked perfetcly OK for 2 days then I've had 3 
> lockups in 20 minutes right after it (a laptop) returned from the lab from 
> "repairs" of its mouse. I'll be investigating that, but the rc5 hangups are 
> in any case a mystery so please help me with any ideas.
> 
> 
> -- 
> 
> 
> Dan Armak
> Gentoo Linux Developer, Desktop Team
> Matan, Israel
> _______________________________________________
> gentoo-dev mailing list
> gentoo-dev@gentoo.org
> http://lists.gentoo.org/mailman/listinfo/gentoo-dev
> 
-- 
Djamil ESSAISSI
Tel:01.58.64.22.44 - Fax:01.58.64.26.81
Systeme Administrateur / Support Technique
www.francexpress.com - www.serveur-express.com

Re: [gentoo-dev] System hangup w/ X taking 100% cpu

[Nathaniel Grady]

To add to this my really old -rc6 system I believe has similar problems - konqueror locks fairly frequently when dragging and dropping (when using as a file manager) - I never ssh'd in to check top because i don't like having any open ports on my laptop. If i ever get my brand-new gentoo install to work with PCMCIA correctly i'll report on that. I do not have this issue with Nautulus or GMC (however i do have a problem that sometmes i get stuck with the "drag and drop" mouse pointer and have to switch to another console to kill -9 gmc or nautulus - i assume that is a backbox or xmms issue). I use default optimization flags for PII in make.conf.

For anyone who is curious this is a Thinkpad 600e / 128mb ram / 366mhz PII / NeoMAGIC video / ambicom ethernet / CS4232 audio w/ oss drivers. W2K did really odd things to the IRQ/DMA assignments but i reflashed the bios and used mkfs to fix that problem :)

--Nate Grady
PS: I'm saying this in hopes that it will help diagnose the problem - I don't really expect to be able to look at much for the next week as it's finals time :( Over xmass break i do plant to get this darn thing workin right :) 

On Thu, Dec 13, 2001 at 07:11:23PM +0200, Dan Armak wrote:
> Hi all,
> 
> On my new r12 system I've had a hangup in X11. No apparent cause. The system 
> stopped accepting any input and the screen didn't even complete a redraw. The 
> mouse went on moving as it was in h/w, but clicks were no good. Magic sysrq 
> wasn't responding either.
> 
> I logged in via ssh, and top showed X was taking 100% cpu. I tried killall X, 
> no effect. I tried kill -9 <X pid>, the remote login hung up and i couldn't 
> ssh to it again. I had to do a hard reboot.
> 
> Then I expirienced the same effect twice on an old rc5 system which underwent 
> absolutely no updates or config changes for a few months now. Only this was 
> standalone not on a net, so I couldn't login via ssh.
> 
> Both rc6 lockups were when running kmail, both rc5 lockups while running 
> ksirc.
> 
> Trying to send this message, I had another lockup on the rc6 system, hope 
> this reaches you.
> 
> I'm absolutely mystified. Ideas? Anyone seen this kind of thing before?
> 
> -- 
> 
> Dan Armak
> Gentoo Linux Developer, Desktop Team
> Matan, Israel
> 
> _______________________________________________
> gentoo-dev mailing list
> gentoo-dev@gentoo.org
> http://lists.gentoo.org/mailman/listinfo/gentoo-dev

Re: [gentoo-dev] System hangup w/ X taking 100% cpu

[Dan Armak]

> Trying to send this message, I had another lockup on the rc6 system, hope
> this reaches you.
I've had a 3rd lockup right after sending this off and  am now working on the 
rc5 machine again. The rc6 one worked perfetcly OK for 2 days then I've had 3 
lockups in 20 minutes right after it (a laptop) returned from the lab from 
"repairs" of its mouse. I'll be investigating that, but the rc5 hangups are 
in any case a mystery so please help me with any ideas.


-- 


Dan Armak
Gentoo Linux Developer, Desktop Team
Matan, Israel

Re: [gentoo-dev] System hangup w/ X taking 100% cpu

[djamil]

just in case this is really, really H/W related, i think i'm gonna
checkout "tux" or something ... a distro made "just for" laptops ...
maybe they have some tricks with these little gizmos...


On Thu, 2001-12-13 at 18:38, Nathaniel Grady wrote:
> To add to this my really old -rc6 system I believe has similar problems - konqueror locks fairly frequently when dragging and dropping (when using as a file manager) - I never ssh'd in to check top because i don't like having any open ports on my laptop. If i ever get my brand-new gentoo install to work with PCMCIA correctly i'll report on that. I do not have this issue with Nautulus or GMC (however i do have a problem that sometmes i get stuck with the "drag and drop" mouse pointer and have to switch to another console to kill -9 gmc or nautulus - i assume that is a backbox or xmms issue). I use default optimization flags for PII in make.conf.
> 
> For anyone who is curious this is a Thinkpad 600e / 128mb ram / 366mhz PII / NeoMAGIC video / ambicom ethernet / CS4232 audio w/ oss drivers. W2K did really odd things to the IRQ/DMA assignments but i reflashed the bios and used mkfs to fix that problem :)
> 
> --Nate Grady
> PS: I'm saying this in hopes that it will help diagnose the problem - I don't really expect to be able to look at much for the next week as it's finals time :( Over xmass break i do plant to get this darn thing workin right :) 
> 
> On Thu, Dec 13, 2001 at 07:11:23PM +0200, Dan Armak wrote:
> > Hi all,
> > 
> > On my new r12 system I've had a hangup in X11. No apparent cause. The system 
> > stopped accepting any input and the screen didn't even complete a redraw. The 
> > mouse went on moving as it was in h/w, but clicks were no good. Magic sysrq 
> > wasn't responding either.
> > 
> > I logged in via ssh, and top showed X was taking 100% cpu. I tried killall X, 
> > no effect. I tried kill -9 <X pid>, the remote login hung up and i couldn't 
> > ssh to it again. I had to do a hard reboot.
> > 
> > Then I expirienced the same effect twice on an old rc5 system which underwent 
> > absolutely no updates or config changes for a few months now. Only this was 
> > standalone not on a net, so I couldn't login via ssh.
> > 
> > Both rc6 lockups were when running kmail, both rc5 lockups while running 
> > ksirc.
> > 
> > Trying to send this message, I had another lockup on the rc6 system, hope 
> > this reaches you.
> > 
> > I'm absolutely mystified. Ideas? Anyone seen this kind of thing before?
> > 
> > -- 
> > 
> > Dan Armak
> > Gentoo Linux Developer, Desktop Team
> > Matan, Israel
> > 
> > _______________________________________________
> > gentoo-dev mailing list
> > gentoo-dev@gentoo.org
> > http://lists.gentoo.org/mailman/listinfo/gentoo-dev
> _______________________________________________
> gentoo-dev mailing list
> gentoo-dev@gentoo.org
> http://lists.gentoo.org/mailman/listinfo/gentoo-dev
> 

Re: [gentoo-dev] System hangup w/ X taking 100% cpu

[Dan Armak]

On Thursday 13 December 2001 19:29, you wrote:
> this is abviously a hardware problem, if you want help, we need to get
> all the specs on your machines (the sick ones) : CPU, RAM MB ...
> intel/amd speed NIC VideoCard.... etc

Toshiba Satellite 3000-400 which includes:
CPU: mobile p3-900
RAM: 256MB
network: 8139too driver (forget the name of the card) (builtin - mini-pci)
Video: nVidia GeForce2Go 16MB RAM with the latest drivers from portage

Nothing in this or anything else seems very helpful, because it worked before.

-- 


Dan Armak
Gentoo Linux Developer, Desktop Team
Matan, Israel

Re: [gentoo-dev] System hangup w/ X taking 100% cpu

[Dan Armak]

It seems this may have something to do with using an external CRT with the 
laptop. It hangs up whenever I try to switch betwen the CRT and the laptop's 
builtin screen. I can't elaborate right now, as I have to go, be back soon. 
Hope that sparcs up some ideas. I may try to downgrade my nvidia driver. Also 
I should start my console in 1024x768 again - haven't set that up yet.


-- 


Dan Armak
Gentoo Linux Developer, Desktop Team
Matan, Israel

Re: [gentoo-dev] System hangup w/ X taking 100% cpu

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 1702 bytes --]

On Thu, 2001-12-13 at 19:11, Dan Armak wrote:
> Hi all,
> 
> On my new r12 system I've had a hangup in X11. No apparent cause. The system 
> stopped accepting any input and the screen didn't even complete a redraw. The 
> mouse went on moving as it was in h/w, but clicks were no good. Magic sysrq 
> wasn't responding either.
> 
> I logged in via ssh, and top showed X was taking 100% cpu. I tried killall X, 
> no effect. I tried kill -9 <X pid>, the remote login hung up and i couldn't 
> ssh to it again. I had to do a hard reboot.
> 
> Then I expirienced the same effect twice on an old rc5 system which underwent 
> absolutely no updates or config changes for a few months now. Only this was 
> standalone not on a net, so I couldn't login via ssh.
> 
> Both rc6 lockups were when running kmail, both rc5 lockups while running 
> ksirc.
> 
> Trying to send this message, I had another lockup on the rc6 system, hope 
> this reaches you.
> 
> I'm absolutely mystified. Ideas? Anyone seen this kind of thing before?
> 
> -- 
> 

Hi Dan

What version kernel, nvidia driver and filesystem are you
running?

I updated my nvidia drivers to the latest 2313, and only had
problems (screen corruption, lockups, etc) on a system that
had been running stable for a long time now. i went back to
the previous version, and it is 100% again.  I also had problems with my
old savage and sertain drivers.

Then, nearly a year back, I had the same type of problems
(lockups), but this was due to bad ram, and a kernel/reiserfs
and hardware compat problem.


Greetings,
-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop Team Developer
Cape Town, South Africa


[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: [gentoo-dev] System hangup w/ X taking 100% cpu

[Dan Armak]

On Thursday 13 December 2001 21:38, you wrote:
>
> Hi Dan
>
> What version kernel, nvidia driver and filesystem are you
> running?
I have kernel 2.4.16 w/ the std patches from portage, nvidia 2314 and 
reiserfs.

>
> I updated my nvidia drivers to the latest 2313, and only had
> problems (screen corruption, lockups, etc) on a system that
> had been running stable for a long time now. i went back to
> the previous version, and it is 100% again.  I also had problems with my
> old savage and sertain drivers.
>
> Then, nearly a year back, I had the same type of problems
> (lockups), but this was due to bad ram, and a kernel/reiserfs
> and hardware compat problem.
Things now seem to have stablizied somewhat after I turned on a 1024x768x16M 
vesa framebuffer for the console. Now it only hangs like that when I press 
Fn+F5, the hotkey for switching between the builtin and external screeens. 
Still I wouldn't call it completely stable.

But I've just found out another possible cause; I compiled agpgart into the 
kernel instead of as a module, and now I can't use NVAGP althuogh I want to. 
So I'll switch that and see if it helps.

-- 

Dan Armak
Gentoo Linux Developer, Desktop Team
Matan, Israel

Re: [gentoo-dev] System hangup w/ X taking 100% cpu

[Dan Armak]

On Thursday 13 December 2001 21:50, you wrote:
> But I've just found out another possible cause; I compiled agpgart into the
> kernel instead of as a module, and now I can't use NVAGP althuogh I want
> to. So I'll switch that and see if it helps.

Well, it didn't help. I may be that the 2.4.14->2.4.16 kernel upgrade is 
causing this. In any case, I'm now starting an rc5->rc6 pugrade of my nat 
router/gateway box, and won't be online for a day or two. Hopefully this'll 
be resolved by then.

-- 

Dan Armak
Gentoo Linux Developer, Desktop Team
Matan, Israel

Re: [gentoo-dev] System hangup w/ X taking 100% cpu

[m3thos]

The same kind of lookups happened to me in my debian box.
I believe that the problem is related with the nvidia drivers, i also use
the latest nvidia driver.


Miguel Sousa Filipe

----- Original Message -----
From: "Dan Armak" <danarmak@gentoo.org>
To: <gentoo-dev@cvs.gentoo.org>
Sent: Thursday, December 13, 2001 5:11 PM
Subject: [gentoo-dev] System hangup w/ X taking 100% cpu


> Hi all,
>
> On my new r12 system I've had a hangup in X11. No apparent cause. The
system
> stopped accepting any input and the screen didn't even complete a redraw.
The
> mouse went on moving as it was in h/w, but clicks were no good. Magic
sysrq
> wasn't responding either.
>
> I logged in via ssh, and top showed X was taking 100% cpu. I tried killall
X,
> no effect. I tried kill -9 <X pid>, the remote login hung up and i
couldn't
> ssh to it again. I had to do a hard reboot.
>
> Then I expirienced the same effect twice on an old rc5 system which
underwent
> absolutely no updates or config changes for a few months now. Only this
was
> standalone not on a net, so I couldn't login via ssh.
>
> Both rc6 lockups were when running kmail, both rc5 lockups while running
> ksirc.
>
> Trying to send this message, I had another lockup on the rc6 system, hope
> this reaches you.
>
> I'm absolutely mystified. Ideas? Anyone seen this kind of thing before?
>
> --
>
> Dan Armak
> Gentoo Linux Developer, Desktop Team
> Matan, Israel
>
> _______________________________________________
> gentoo-dev mailing list
> gentoo-dev@gentoo.org
> http://lists.gentoo.org/mailman/listinfo/gentoo-dev

Re: [gentoo-dev] System hangup w/ X taking 100% cpu

[Dan Armak]

Well, to close this, things seem to have stabilized more or less. The only 
thing that hangs the system without fail is swtching via the Fn+F5 hotkey 
between the builin and external monitors.

I'll try to downgrade the kernel to 2.4.14 later and see if that helps, or I 
may just wait for 2.4.17 (which has lots of fixes I need in 2 drivers) and 
see if that works.

Thanks, everyone, for your help.

-- 

Dan Armak
Gentoo Linux Developer, Desktop Team
Matan, Israel