Mailing List Archive

On Wed, 2004-10-06 at 11:18, Kevin Philp wrote:
> I have been using Linux for several years and recently installed Gentoo
> 2004.2. I run the system on stable packages (ie no ~x86) and it runs great
> with the exception of occasional system freezes. They don't sem to be related
> to any particular package, one day it freeze on Gnucash, the next it freezes
> on firebird. The system has a VIA chipset based motherboard, NVidia graphics
> card (only uses nv driver), AMD XP1800 and I am using the KDE desktop. Anyone
> any tips or hints what might be going on? I haven't seen this behaviour since
> dumping Win95!

In my experience random program freezes have usually meant faulty
memory. Try testing your memory with a program such as memtest86( it's
in portage ) and see what the results are before trying to diagnose the
problem any further.

Regards,
--
Nelis Lamprecht
ICTS
University of Cape Town


--
gentoo-user@gentoo.org mailing list

On Wed, 2004-10-06 at 04:18, Kevin Philp wrote:
> I have been using Linux for several years and recently installed Gentoo
> 2004.2. I run the system on stable packages (ie no ~x86) and it runs great
> with the exception of occasional system freezes. They don't sem to be related
> to any particular package, one day it freeze on Gnucash, the next it freezes
> on firebird.

I'd recommend letting memtest86 run overnight. That sounds just like a
ram problem I had a while back.
--
Owen Ford <oford@arghblech.com>

() ascii ribbon campaign - against html e-mail
/\ - against proprietary attachments

Kevin,

On Wed, October 6, 2004 11:18, Kevin Philp said:
> I have been using Linux for several years and recently installed Gentoo
> 2004.2. I run the system on stable packages (ie no ~x86) and it runs great
> with the exception of occasional system freezes. They don't sem to be
> related
> to any particular package, one day it freeze on Gnucash, the next it
> freezes
> on firebird. The system has a VIA chipset based motherboard, NVidia
> graphics
> card (only uses nv driver), AMD XP1800 and I am using the KDE desktop.
> Anyone
> any tips or hints what might be going on? I haven't seen this behaviour
> since
> dumping Win95!

it depends on the kernel you are using.
I have the latest 2.6-mm kernel and since a week I also experience random
crashes. it seems like the anticipatory io-schedular does not work as it
should and when I boot with elevator=cfq on the kernel-commandline all
works well.
maybe you used a troublesome set of CFLAGS or a wrong CHOST?


Rudmer



--
gentoo-user@gentoo.org mailing list

>
> I'd recommend letting memtest86 run overnight.

I have a similar problem and would like to ask: is there some level of
acceptable number of errors after an overnight run of memtest or
should the list be completely clear on a non-faulty hardware?

--
gentoo-user@gentoo.org mailing list

On Wed, Oct 06, 2004 at 12:47:16PM +0200, Peter Ga??i wrote:
> I have a similar problem and would like to ask: is there some level of
> acceptable number of errors after an overnight run of memtest or
> should the list be completely clear on a non-faulty hardware?

On any good hardware I've had in the past, memtest reported zero errors.
Tom

On 12:41 Wed 06 Oct , Nelis Lamprecht wrote:
> On Wed, 2004-10-06 at 11:18, Kevin Philp wrote:
> > I have been using Linux for several years and recently installed Gentoo
> > 2004.2. I run the system on stable packages (ie no ~x86) and it runs great
> > with the exception of occasional system freezes. They don't sem to be related
> > to any particular package, one day it freeze on Gnucash, the next it freezes
> > on firebird. The system has a VIA chipset based motherboard, NVidia graphics
> > card (only uses nv driver), AMD XP1800 and I am using the KDE desktop. Anyone
> > any tips or hints what might be going on? I haven't seen this behaviour since
> > dumping Win95!
>
> In my experience random program freezes have usually meant faulty
> memory. Try testing your memory with a program such as memtest86( it's
> in portage ) and see what the results are before trying to diagnose the
> problem any further.

You might also try pulling out all of your expansion cards and memory and
re-seating them. They'll work their way loose over time, especially if you
boot your computer often.

Bill

--
gentoo-user@gentoo.org mailing list

Thomas Kirchner wrote:

>On Wed, Oct 06, 2004 at 12:47:16PM +0200, Peter Ga??i wrote:
>
>
>>I have a similar problem and would like to ask: is there some level of
>>acceptable number of errors after an overnight run of memtest or
>>should the list be completely clear on a non-faulty hardware?
>>
>>
>
>On any good hardware I've had in the past, memtest reported zero errors.
>Tom
>
>
One other thing that you should look at is the speed that your ram is
set at. Go into your BIOS and manually set the speed to a lower value
and run memtest again. Not all BIOS' detect the correct max speed of
your RAM.

--
Ted Ozolins(VE7TVO)
Westbank, B. C


--
gentoo-user@gentoo.org mailing list

On Wednesday 06 October 2004 11:18, Kevin Philp wrote:
> I have been using Linux for several years and recently installed Gentoo
> 2004.2. I run the system on stable packages (ie no ~x86) and it runs great
> with the exception of occasional system freezes. They don't sem to be
> related to any particular package, one day it freeze on Gnucash, the next
> it freezes on firebird. The system has a VIA chipset based motherboard,
> NVidia graphics card (only uses nv driver), AMD XP1800 and I am using the
> KDE desktop. Anyone any tips or hints what might be going on? I haven't
> seen this behaviour since dumping Win95!

if you can rule out defective ram/bad bios settings, try a new, powerfull,
branded PSU (like enermax), do not buy noname-PSU, don't buy 'pseudo' brands
like Fortron.
A lot of people never check their PSU and would be shocked how fast this beast
get 'old' and start to under/overvoltage or letting all the 'noise' of the
power lines hit the board, with very bad results: your system will become
totally instable, up to the moment, where starting a neonlamp will crash your
system. Without a good power source, all signals in your box will suffer,
your cpu does not get enough ampere when it needs them and get 'confused and
so on...

--
Conclusions
In a straight-up fight, the Empire squashes the Federation like a bug. Even
with its numerical advantage removed, the Empire would still squash the
Federation like a bug. Accept it. -Michael Wong

--
gentoo-user@gentoo.org mailing list

Well I ran memtest for several hours and everything was OK with the memory. I
wiggled and pushed all the boards back into their slots but they didn't seem
to need it, I added append="noapic acpi=off" to lilo, I checked the IRQ's but
don't see any over lapping problems, I checked the cpu temperature and its a
nice cool 39C... I am running out of ideas! I mainly get the problem when the
machine has been left for several hours unattended, moving the mouse brings
the screen back up but everything is frozen solid and needs a full reboot.

Scanning through the forums random system freezes seems to be quite a common
problem, but I have yet to see a definitive answer...if I find one I will
post it in big bold letters!

The machine can also be booted into Mandrake. I will leave it on Mandrake
tonight to see if I get the same problem or wether its a Gentoo problem. If I
do see the same problem then I will look at the power supply.

Kevin.


On Wednesday 6 October 2004 19:21, Hemmann, Volker Armin wrote:
On Wednesday 06 October 2004 11:18, Kevin Philp wrote:
> I have been using Linux for several years and recently installed Gentoo
> 2004.2. I run the system on stable packages (ie no ~x86) and it runs great
> with the exception of occasional system freezes. They don't sem to be
> related to any particular package, one day it freeze on Gnucash, the next
> it freezes on firebird. The system has a VIA chipset based motherboard,
> NVidia graphics card (only uses nv driver), AMD XP1800 and I am using the
> KDE desktop. Anyone any tips or hints what might be going on? I haven't
> seen this behaviour since dumping Win95!

if you can rule out defective ram/bad bios settings, try a new, powerfull,
branded PSU (like enermax), do not buy noname-PSU, don't buy 'pseudo' brands
like Fortron.
A lot of people never check their PSU and would be shocked how fast this
beast get 'old' and start to under/overvoltage or letting all the 'noise' of
the power lines hit the board, with very bad results: your system will
become totally instable, up to the moment, where starting a neonlamp will
crash your system. Without a good power source, all signals in your box will
suffer, your cpu does not get enough ampere when it needs them and get
'confused and so on...


--
gentoo-user@gentoo.org mailing list

Kevin Philp wrote:
> Well I ran memtest for several hours and everything was OK with the memory. I
> wiggled and pushed all the boards back into their slots but they didn't seem
> to need it, I added append="noapic acpi=off" to lilo, I checked the IRQ's but
> don't see any over lapping problems, I checked the cpu temperature and its a
> nice cool 39C... I am running out of ideas! I mainly get the problem when the
> machine has been left for several hours unattended, moving the mouse brings
> the screen back up but everything is frozen solid and needs a full reboot.
>
> Scanning through the forums random system freezes seems to be quite a common
> problem, but I have yet to see a definitive answer...if I find one I will
> post it in big bold letters!
>
> The machine can also be booted into Mandrake. I will leave it on Mandrake
> tonight to see if I get the same problem or wether its a Gentoo problem. If I
> do see the same problem then I will look at the power supply.

Check your graphics card and X configuration as well. My system would
freeze at random until I removed "AGPMode" "4" from /etc/X11/xorg.conf.
I have a Matrox G450, which I know is not what you have, but it could be
worth a try.

-Arnstein

--
gentoo-user@gentoo.org mailing list

Kevin Philp wrote:

> I have been using Linux for several years and recently installed Gentoo
> 2004.2. I run the system on stable packages (ie no ~x86) and it runs great
> with the exception of occasional system freezes. They don't sem to be
> related to any particular package, one day it freeze on Gnucash, the next
> it freezes on firebird. The system has a VIA chipset based motherboard,
> NVidia graphics card (only uses nv driver), AMD XP1800 and I am using the
> KDE desktop. Anyone any tips or hints what might be going on? I haven't
> seen this behaviour since dumping Win95!

Another thing to try might be using the 'nvidia' driver instead of the 'nv'
one. I know my experiences don't mesh with a lot of people, but I always
found the nvidia driver to be much more stable and reliable than 'nv'.
Your mileage may vary, but it could also be with a shot while you're
experimenting.


--
gentoo-user@gentoo.org mailing list

On Sun, Oct 10, 2004 at 02:25:28PM +0100, Kevin Philp wrote:
> Well I ran memtest for several hours and everything was OK with the memory. I
> wiggled and pushed all the boards back into their slots but they didn't seem
> to need it, I added append="noapic acpi=off" to lilo, I checked the IRQ's but
> don't see any over lapping problems, I checked the cpu temperature and its a
> nice cool 39C... I am running out of ideas! I mainly get the problem when the
> machine has been left for several hours unattended, moving the mouse brings
> the screen back up but everything is frozen solid and needs a full reboot.
>
> Scanning through the forums random system freezes seems to be quite a common
> problem, but I have yet to see a definitive answer...if I find one I will
> post it in big bold letters!

That's because there isn't one ultimate solution. The symptoms may be
the same but the causes can vary.

Hardware wise i'd look at the powersupply. Replace it temporarly with a
spare and test. Some psu's drop their voltages at random intervals. The
problem is that psus have become so cheap that their quality is simply
lousy.

Also emerge cpuburn and try to set your system under high load and see
if it lives after a couple of hours.

Also check the fan of the vga board. If it doesn't turn it will freeze
(graphics only, system is accessable via ssh) on any more power
intensive graphics operation.

Softwarewise i'd try taking out agpgart support from the kernel. In the
past I had a mirad of issues with it, although never had one with the
2.6 series.

So.. more shots in the dark. Unstable systems are always tricky.

- Folken

Its now solved.

It mainly kicked in after the machine had gone into suspend mode. I checked
memory, temperatures and It was finally solved by adding

append="noapic nolapic acpi=off"

The machine has now worked for about 2 weeks/24 hours a day without a problem.

Kevin.


On Tuesday 26 October 2004 08:32, Folken wrote:
On Sun, Oct 10, 2004 at 02:25:28PM +0100, Kevin Philp wrote:
> Well I ran memtest for several hours and everything was OK with the memory.
> I wiggled and pushed all the boards back into their slots but they didn't
> seem to need it, I added append="noapic acpi=off" to lilo, I checked the
> IRQ's but don't see any over lapping problems, I checked the cpu
> temperature and its a nice cool 39C... I am running out of ideas! I mainly
> get the problem when the machine has been left for several hours
> unattended, moving the mouse brings the screen back up but everything is
> frozen solid and needs a full reboot.
>
> Scanning through the forums random system freezes seems to be quite a
> common problem, but I have yet to see a definitive answer...if I find one I
> will post it in big bold letters!

That's because there isn't one ultimate solution. The symptoms may be
the same but the causes can vary.

Hardware wise i'd look at the powersupply. Replace it temporarly with a
spare and test. Some psu's drop their voltages at random intervals. The
problem is that psus have become so cheap that their quality is simply
lousy.

Also emerge cpuburn and try to set your system under high load and see
if it lives after a couple of hours.

Also check the fan of the vga board. If it doesn't turn it will freeze
(graphics only, system is accessable via ssh) on any more power
intensive graphics operation.

Softwarewise i'd try taking out agpgart support from the kernel. In the
past I had a mirad of issues with it, although never had one with the
2.6 series.

So.. more shots in the dark. Unstable systems are always tricky.

- Folken


--
gentoo-user@gentoo.org mailing list