|
Next: SUSE: Internet connection dies/won't start
|
| Author |
Message |
External
Since: Nov 29, 2004
Posts: 71
|
(Msg. 1) Posted: Wed Feb 06, 2008 7:48 am
Post subject: Almost random system freezes - SUSE10.3 - Diagnostic Help Needed
Archived from groups: alt>os>linux>suse (more info?)
|
|
|
Hi, Guys
Been happy with 10.3 since I installed the RT kernel from the jengelh
repository, and it has been running perfectly well for several months.
I've been updating it regularly, but deliberately did not touch either
the kernel or the NVidia driver.
Suddenly, about 3 weeks ago my system locked up about 5 minutes after
logging in. I rebooted, and the system was fine. Ever since then, this
has happened on a fairly regular basis - practically, but not absolutely
every day.
I've looked at (what I consider) the obvious:
- There are no entries in the syslog associated with the crash.
- It happens both when I'm using the system, or if I just log
in to the desktop and do nothing.
- It happens whether ACPI and APM are disabled or not.
- It happens with both the default and RT kernels.
- I updated the NVidia driver - still happens.
- I updated both default and RT kernels - still happens.
- I've changed desktop (was KDE, now XFCE) - no change.
The lock-up only happens when I'm logged in. I can leave the system
running for an hour before I log in, and it still happens.
I haven't tried creating another user yet, so I don't know if it's
account specific.
Can anyone think of something else to check? I am running out of ideas,
and considering a rebuild. Many thanks in advance.
--
David Haggett
Linux user since 01/01/2003
Email: david<at>haggett<dot>demon<dot>co<dot>uk |
|
| Back to top |
|
 | |
External
Since: Nov 05, 2007
Posts: 16
|
(Msg. 2) Posted: Wed Feb 06, 2008 1:02 pm
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
On Wed, 2008-02-06 at 07:48 +0000, David Haggett wrote:
> Hi, Guys
>
> Been happy with 10.3 since I installed the RT kernel from the jengelh
> repository, and it has been running perfectly well for several months.
> I've been updating it regularly, but deliberately did not touch either
> the kernel or the NVidia driver.
....
> - I updated the NVidia driver - still happens.
You might want to try using an older Nvidia driver (potentially
much older). The last several ones from Nvidia all have
problems. If you can use nv, use it instead.
Are you updating the Nvidia driver via rpm or from rebuild using
the .sh script from Nvidia? (just asking out of curiosity) |
|
| Back to top |
|
| |
External
Since: Nov 10, 2007
Posts: 6
|
(Msg. 3) Posted: Wed Feb 06, 2008 7:18 pm
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help Needed [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
David Haggett wrote:
> Been happy with 10.3 since I installed the RT kernel from the jengelh
> repository, and it has been running perfectly well for several months.
> I've been updating it regularly, but deliberately did not touch either
> the kernel or the NVidia driver.
I had a similar experience. It would freeze and I had to reboot.
I gave up and will wait until there is some newer version.
I think that this is something to do with the video drivers.
I tried the rt kernel from suse, but the nvidia driver would not compile (!)
and complain that the sources were not installed (they were).
I was quite happy with suse 10.2 and the jacklab kernel but my experience
with the jengelh rt-kernel was not as happy, unfortunately.
Have you tried to talk to the jacklab guys in the irc channel? They have
been very friendly and gave good advice when installing the jacklab kernel
in suse 10.2.
Please post back if you manage to solve it.
Cheers,
-G- |
|
| Back to top |
|
| |
External
Since: Nov 29, 2004
Posts: 71
|
(Msg. 4) Posted: Thu Feb 07, 2008 7:54 am
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
On Wed, 06 Feb 2008 13:02:35 -0600, Chris Cox wrote:
> On Wed, 2008-02-06 at 07:48 +0000, David Haggett wrote:
>> Hi, Guys
>>
>> Been happy with 10.3 since I installed the RT kernel from the jengelh
>> repository, and it has been running perfectly well for several months.
>> I've been updating it regularly, but deliberately did not touch either
>> the kernel or the NVidia driver.
> ...
>> - I updated the NVidia driver - still happens.
>
> You might want to try using an older Nvidia driver (potentially much
> older). The last several ones from Nvidia all have problems. If you
> can use nv, use it instead.
Just to prove my earlier analysis wrong, last night the PC froze at the
KDM screen after being on around 5 minutes (I think, I was in the shower)
and this morning X just reloaded while I was browsing in Firefox.
I will take this advice. I have 100.14.11 still on my hard-drive, which
I works with my GeForce 7600 GS.
> Are you updating the NVidia driver via rpm or from rebuild using the .sh
> script from NVidia? (just asking out of curiosity)
Always the NVidia script. I tried the RPMs once, but couldn't be
bothered to work out what I needed to do after they were installed. The
script just works.
--
David Haggett
Linux user since 01/01/2003
Email: david<at>haggett<dot>demon<dot>co<dot>uk |
|
| Back to top |
|
| |
External
Since: Nov 29, 2004
Posts: 71
|
(Msg. 5) Posted: Fri Feb 08, 2008 8:23 am
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
On Thu, 07 Feb 2008 10:57:49 +0000, Baron wrote:
>> The PSU is actually pretty new, as is the graphics card (both replaced
>> for noise reasons), but the rest of the computer is probably 3 years
>> old based on Linux Format's 2004 Ultimate Linux Box (AMD64 3200+ on a
>> MSI K8T Neo FS12R main board).
>>
>> I'll take the side off and have a look at the w/e.
>
> The PSU that I was referring to is not the big metal box, but the ones
> on the mainboard. In my case, probably yours as well, there are at
> least two. One is for the CPU and the other is for the memory.
Thanks for the clarification. I think I know which ones they are.
Certainly recognise the CPU coil, although I don't recall it getting that
hot on my system (not that I'm in the habit of touching internal
components while they are switched on.
It strikes me as strange that it should be a hardware fault because once
it gets past that first 5 minutes or so it just runs for ever (well it's
a workstation so it normally only stays up for 18 hours at a time).
Then again, I suppose capacitor failures are most likely when they are
either charging or discharging. I've not ruled out the graphics card
driver yet.
I think I should start it in text mode and see what happens a few times.
--
David Haggett
Linux user since 01/01/2003
Email: david<at>haggett<dot>demon<dot>co<dot>uk |
|
| Back to top |
|
| |
External
Since: Jan 25, 2008
Posts: 15
|
(Msg. 6) Posted: Fri Feb 08, 2008 12:49 pm
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help Needed [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
David Haggett wrote:
> On Thu, 07 Feb 2008 10:57:49 +0000, Baron wrote:
>
>>> The PSU is actually pretty new, as is the graphics card (both
>>> replaced for noise reasons), but the rest of the computer is
>>> probably 3 years old based on Linux Format's 2004 Ultimate Linux Box
>>> (AMD64 3200+ on a MSI K8T Neo FS12R main board).
>>>
>>> I'll take the side off and have a look at the w/e.
>>
>> The PSU that I was referring to is not the big metal box, but the
>> ones on the mainboard. In my case, probably yours as well, there are
>> at least two. One is for the CPU and the other is for the memory.
>
> Thanks for the clarification. I think I know which ones they are.
> Certainly recognise the CPU coil, although I don't recall it getting
> that hot on my system (not that I'm in the habit of touching internal
> components while they are switched on.
As long as you remember to make sure that your body and the machine case
are the same potential ie hang on to the earthed metal case with one
hand, there is no danger to you! Other than a burnt finger 
Seriously though not all run hot. A thermistor bead probe helps to
quantify the actual temperatures.
> It strikes me as strange that it should be a hardware fault because
> once it gets past that first 5 minutes or so it just runs for ever
> (well it's a workstation so it normally only stays up for 18 hours at
> a time).
I see a lot of machines that behave in a similar way! There are other
causes, its just that 75% turn out to be hardware related.
I had one last week. Hard drive not spinning up every time. Turned out
to be a broken power plug pin. Not at all obvious at first. In fact I
didn't find it until I put the old drive into a test machine and
discovered that the power plug wouldn't fit !!
> Then again, I suppose capacitor failures are most likely when they are
> either charging or discharging. I've not ruled out the graphics card
> driver yet.
>
> I think I should start it in text mode and see what happens a few
> times.
If its a higher end card they very often have their own SMPS on the
card! Bad caps here can cause the problem you describe.
--
Regards:
Baron. |
|
| Back to top |
|
| |
External
Since: Nov 29, 2004
Posts: 71
|
(Msg. 7) Posted: Fri Feb 08, 2008 10:12 pm
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
On Fri, 08 Feb 2008 12:49:01 +0000, Baron wrote:
>> Then again, I suppose capacitor failures are most likely when they are
>> either charging or discharging. I've not ruled out the graphics card
>> driver yet.
>>
>> I think I should start it in text mode and see what happens a few
>> times.
Out of a total of 1 start to run-level 3, it crashed 0 times.
>
> If its a higher end card they very often have their own SMPS on the
> card! Bad caps here can cause the problem you describe.
It's a GeForce 7600 GS so about as high-end as you can get in consumer
AGP. Uses a molex connector, which an implies on-board power system. I
chose it because it is fanless, but the fact its about 3x to 6x faster
than my previous card and cost half the price is a bit of a bonus also.
--
David Haggett
Linux user since 01/01/2003
Email: david<at>haggett<dot>demon<dot>co<dot>uk |
|
| Back to top |
|
| |
External
Since: Feb 13, 2007
Posts: 25
|
(Msg. 8) Posted: Fri Feb 08, 2008 10:52 pm
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help Needed [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
David Haggett wrote:
> On Fri, 08 Feb 2008 12:49:01 +0000, Baron wrote:
>
>>> Then again, I suppose capacitor failures are most likely when they
>>> are either charging or discharging. I've not ruled out the graphics
>>> card driver yet.
>>>
>>> I think I should start it in text mode and see what happens a few
>>> times.
>
> Out of a total of 1 start to run-level 3, it crashed 0 times.
Run level 3 is text based so its won't stress the graphics card a great
deal.
>> If its a higher end card they very often have their own SMPS on the
>> card! Bad caps here can cause the problem you describe.
>
> It's a GeForce 7600 GS so about as high-end as you can get in consumer
> AGP. Uses a molex connector, which an implies on-board power system.
> I chose it because it is fanless, but the fact its about 3x to 6x
> faster than my previous card and cost half the price is a bit of a
> bonus also.
Ok! Two suggestions. Have a look in the BIOS and see if you can reduce
the video level as low as it will go and swap back your old card.
But I must confess the more I think about it I don't think your problem
has anything to do with video.
I keep an old PCI video card around for testing. Its surprising how
many on-board video failures you can solve using one.
--
Best Regards:
Baron. |
|
| Back to top |
|
| |
External
Since: Nov 29, 2004
Posts: 71
|
(Msg. 9) Posted: Sat Feb 09, 2008 9:52 am
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
On Fri, 08 Feb 2008 22:52:45 +0000, Baron wrote:
>> Out of a total of 1 start to run-level 3, it crashed 0 times.
>
> Run level 3 is text based so its won't stress the graphics card a great
> deal.
Sorry, I should have clarified. I started in text-mode, let it run for
10 mins or so while I was in the shower, then logged in as root and did
"telinit 5 && exit".
Did the same this morning and again no crash. I will have to continue
the experiment for a few more days, as whenever I think I've made
headway, the machine proves me wrong.
Also is interesting (to me) that if I just reboot after a crash using the
reset button, the PC doesn't freeze again. Could it be something to do
with the state the OS is at when X starts maybe?
--
David Haggett
Linux user since 01/01/2003
Email: david<at>haggett<dot>demon<dot>co<dot>uk |
|
| Back to top |
|
| |
External
Since: Jun 05, 2007
Posts: 86
|
(Msg. 10) Posted: Sat Feb 09, 2008 10:33 am
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help Needed [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
On 2008-02-09, David Haggett <news-spam DeleteThis @haggett.demon.co.uk> wrote:
> On Fri, 08 Feb 2008 22:52:45 +0000, Baron wrote:
>
>>> Out of a total of 1 start to run-level 3, it crashed 0 times.
>>
>> Run level 3 is text based so its won't stress the graphics card a great
>> deal.
>
> Sorry, I should have clarified. I started in text-mode, let it run for
> 10 mins or so while I was in the shower, then logged in as root and did
> "telinit 5 && exit".
>
> Did the same this morning and again no crash. I will have to continue
> the experiment for a few more days, as whenever I think I've made
> headway, the machine proves me wrong.
>
> Also is interesting (to me) that if I just reboot after a crash using the
> reset button, the PC doesn't freeze again. Could it be something to do
> with the state the OS is at when X starts maybe?
>
>
I'm just putting my 2 cents here.
I have had similar symptoms as you describe.
I've also then had the behavior you describe, after being able to run the
machine for some time it will run.
It's always been swollen capacitors. Either on the MOBO, most of the
time, or the power supply.
I'm betting for the mother board to have at least one or two maybe
slightly swollen capacitors. Make sure you don't have this
Maybe also the video card, although mine have not yet been there.
I think that this "swollen capacitor decease" has become more and more
general. Probably due to cheaper and cheaper building of components.
All my oldest hardware is without this problem, it's on the newer stuff.
(Newer in my case is not new, but older is old).
Because of this phenomenom I'm going to invest into a proper server
mother board in the future, I'm hearing that as they are built from
proper components they don't suffer so much from this "cheepnis".
Vahis
Remodeling my site to train new things:
http://waxborg.servepics.com
--
"The only thing more expensive than training is the lack of it"
Henry Ford |
|
| Back to top |
|
| |
External
Since: Feb 13, 2007
Posts: 25
|
(Msg. 11) Posted: Sat Feb 09, 2008 1:49 pm
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help Needed [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
David Haggett wrote:
> On Fri, 08 Feb 2008 22:52:45 +0000, Baron wrote:
>
>>> Out of a total of 1 start to run-level 3, it crashed 0 times.
>>
>> Run level 3 is text based so its won't stress the graphics card a
>> great deal.
>
> Sorry, I should have clarified. I started in text-mode, let it run
> for 10 mins or so while I was in the shower, then logged in as root
> and did "telinit 5 && exit".
>
> Did the same this morning and again no crash. I will have to continue
> the experiment for a few more days, as whenever I think I've made
> headway, the machine proves me wrong.
Can you boot from a live CD and run your machine for a while ? That
will rule out your software if you still get the problem.
> Also is interesting (to me) that if I just reboot after a crash using
> the reset button, the PC doesn't freeze again. Could it be something
> to do with the state the OS is at when X starts maybe?
>
Hi David,
I agree that is interesting! Pressing reset doesn't cause the power
supplies to stop and restart, so at that point they would be stable.
--
Best Regards:
Baron. |
|
| Back to top |
|
| |
External
Since: Feb 13, 2007
Posts: 25
|
(Msg. 12) Posted: Sat Feb 09, 2008 2:03 pm
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help Needed [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
Vahis wrote:
> On 2008-02-09, David Haggett <news-spam.DeleteThis@haggett.demon.co.uk> wrote:
>> On Fri, 08 Feb 2008 22:52:45 +0000, Baron wrote:
>>
>>>> Out of a total of 1 start to run-level 3, it crashed 0 times.
>>>
>>> Run level 3 is text based so its won't stress the graphics card a
>>> great deal.
>>
>> Sorry, I should have clarified. I started in text-mode, let it run
>> for 10 mins or so while I was in the shower, then logged in as root
>> and did "telinit 5 && exit".
>>
>> Did the same this morning and again no crash. I will have to
>> continue the experiment for a few more days, as whenever I think I've
>> made headway, the machine proves me wrong.
>>
>> Also is interesting (to me) that if I just reboot after a crash using
>> the
>> reset button, the PC doesn't freeze again. Could it be something to
>> do with the state the OS is at when X starts maybe?
>>
>>
>
> I'm just putting my 2 cents here.
> I have had similar symptoms as you describe.
> I've also then had the behavior you describe, after being able to run
> the machine for some time it will run.
>
> It's always been swollen capacitors. Either on the MOBO, most of the
> time, or the power supply.
>
> I'm betting for the mother board to have at least one or two maybe
> slightly swollen capacitors. Make sure you don't have this
>
> Maybe also the video card, although mine have not yet been there.
>
> I think that this "swollen capacitor decease" has become more and more
> general. Probably due to cheaper and cheaper building of components.
>
> All my oldest hardware is without this problem, it's on the newer
> stuff. (Newer in my case is not new, but older is old).
>
> Because of this phenomenom I'm going to invest into a proper server
> mother board in the future, I'm hearing that as they are built from
> proper components they don't suffer so much from this "cheepnis".
>
>
> Vahis
>
> Remodeling my site to train new things:
> http://waxborg.servepics.com
Hi Vahis,
I'm of a similar opinion! Part of the problem seems to be on the newer
boards there may not be any physical signs that the capacitors have
gone bad. I look at the temperature of suspect devices, for instance
if a capacitor temperature continues to rise over a long period then it
would be suspect. A good cap's temperature would rise and then
stabilise at some level over a fairly short period without exceeding
its ratings, usually 105C.
A cheap thermistor bead probe is very useful in determining these
temperatures. I use a short piece of rubber tube sleeving to trap it
against the suspect device.
--
Best Regards:
Baron. |
|
| Back to top |
|
| |
External
Since: Nov 29, 2004
Posts: 71
|
(Msg. 13) Posted: Sat Feb 09, 2008 4:35 pm
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
On Sat, 09 Feb 2008 13:49:37 +0000, Baron wrote:
>> Also is interesting (to me) that if I just reboot after a crash using
>> the reset button, the PC doesn't freeze again. Could it be something
>> to do with the state the OS is at when X starts maybe?
> I agree that is interesting! Pressing reset doesn't cause the power
> supplies to stop and restart, so at that point they would be stable.
Ah! Again pointing to the possibility of a capacitor failure.
Can you tell that I *really* don't want to take the PC apart to fix
this? Still I suppose it then becomes possible to get a new MOBO with
SATA2 and support for more RAM (if I can find one that still supports my
CPU).
--
David Haggett
Linux user since 01/01/2003
Email: david<at>haggett<dot>demon<dot>co<dot>uk |
|
| Back to top |
|
| |
External
Since: Feb 13, 2007
Posts: 25
|
(Msg. 14) Posted: Sat Feb 09, 2008 5:37 pm
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help Needed [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
David Haggett wrote:
> On Sat, 09 Feb 2008 13:49:37 +0000, Baron wrote:
>
>>> Also is interesting (to me) that if I just reboot after a crash
>>> using the reset button, the PC doesn't freeze again. Could it be
>>> something to do with the state the OS is at when X starts maybe?
>
>
>> I agree that is interesting! Pressing reset doesn't cause the power
>> supplies to stop and restart, so at that point they would be stable.
>
> Ah! Again pointing to the possibility of a capacitor failure.
>
> Can you tell that I *really* don't want to take the PC apart to fix
> this?
I don't remember.... Did you say you had taken the side off and had a
look ?
> Still I suppose it then becomes possible to get a new MOBO with
> SATA2 and support for more RAM (if I can find one that still supports
> my CPU).
The "New PC" is one argument I don't want to get into! Her indoors
always wins !!!
First easy option. Try a new PSU, borrow one from another machine if
you have to. Pity you are a bit too far away to help directly.
--
Best Regards:
Baron. |
|
| Back to top |
|
| |
External
Since: Nov 29, 2004
Posts: 71
|
(Msg. 15) Posted: Sat Feb 09, 2008 6:45 pm
Post subject: Re: Almost random system freezes - SUSE10.3 - Diagnostic Help [Login to view extended thread Info.]
Archived from groups: per prev. post (more info?)
|
|
|
On Sat, 09 Feb 2008 17:37:57 +0000, Baron wrote:
>> Can you tell that I *really* don't want to take the PC apart to fix
>> this?
>
> I don't remember.... Did you say you had taken the side off and had a
> look ?
I have now, and to be honest, apart from a bit of dust everything looks
perfectly normal.
I think I've identified the PSU capacitors (right next to the power
connector, just above the coil , quite large with a green casing). BTW
the coil is not hot to the touch.
Not sure about the RAM ones, but none of the components on the board
appear to be swollen. There is one (in a row of 8 next to the CPU, just
behind the edge of the heat-sink that isn't on the board straight, but it
never was, even when I built the PC.
> The "New PC" is one argument I don't want to get into! Her indoors
> always wins !!!
Similar story here - complicated by a love of guitars (no you're not
getting a Strat to go with the ES335!). In fact the PC has now become
just another component of my music room - my amp, effects, metronome,
synth, HDR and mixing desk.
> First easy option. Try a new PSU, borrow one from another machine if
> you have to. Pity you are a bit too far away to help directly.
You may be right - although as I said, this one isn't old (about 6
months) and wasn't cheap either. Out of interest, where are you?
--
David Haggett
Linux user since 01/01/2003
Email: david<at>haggett<dot>demon<dot>co<dot>uk |
|
| Back to top |
|
| |
|
You can post new topics in this forum
You can reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
FAQ
Search
Profile
Private Messages
Log in/Register/Password
Linux Forums
RSS | Data
Feeds | Advertise | Submit | Bookmark | Newsletter | Contact