The machine on which I am typing this is an IBM Netvista (8318-31G)
with an ATI Radeon 7000 PCI graphics card, running Ubuntu 8.04.

Sometimes it runs for days or even a week without a problem - but it
also has a tendency to freeze utterly. No response to keyboard or
mouse, won't accept ssh, not reachable by ping, no response to Alt
+SysRq. Nothing in /var/log/syslog. Nothing obvious the Xorg log
either, though the most recent crashed one has this at the end

Entering Restore TV
Restore TV PLL
Restore TVHV
Restore TV Restarts
Restore Timing Tables
Restore TV standard
Leaving Restore TV

and the current one doesn't. I can't track this down to any particular
programme - twice this week it has done it when left at the login
screen with noone logged on.

Any bright ideas? The problem seems to have started with 8.04: I never
had this with 7.10.

Ian

On 2008-07-10, Ian  wrote:

> Any bright ideas? The problem seems to have started with 8.04: I never
> had this with 7.10.

Well, first off, do a memory test, don't pay too much attention to the
change from 7.10 just yet.  Also, check for excessive heat, I've had
problems in the past with machines overheating due to ACPI problems.
Another strange change I had was my main hard disc powering down on a
machine, only after upgrading a kernel, the new kernel had the smarts
to power the drive down, but sadly not enough to power it back up
again, so watch for that too ;-)

Just some random stuff of the top of my head, easy to check for, go
for the easy stuff first ;-)

It does sound like a CPU locking up due to heat, but that's just a
vague feeling wafting across a text-only medium so don't attach too
much weight to it..

-- 
Blast off and strike the evil Bydo empire!
http://youtube.com/user/tarcus69
http://www.flickr.com/photos/tarcus/sets/

In article ,
	Ian Rawlings wrote:

> It does sound like a CPU locking up due to heat, but that's just a
> vague feeling wafting across a text-only medium so don't attach too
> much weight to it..

I'd look to that, too, given that the ambient temperature is raised a
bit at the moment.

Run memtest overnight. Then boot up in a single user mode and run
programs from cpuburn (burnBX and burnP6) for a few minutes each.

-- 
Paul Martin

In ,
Paul Martin  wrote:

> In article ,
> 	Ian Rawlings wrote:
>
>> It does sound like a CPU locking up due to heat, but that's just a
>> vague feeling wafting across a text-only medium so don't attach too
>> much weight to it..
>
> I'd look to that, too, given that the ambient temperature is raised a
> bit at the moment.
>
> Run memtest overnight. Then boot up in a single user mode and run
> programs from cpuburn (burnBX and burnP6) for a few minutes each.

See if you can monitor the temperature with lm-sensors and one of its
front-ends. If not, reboot when everything's warmed up and check it in
the BIOS.

-- 
TH * http://www.realh.co.uk

Ian wrote:
> The machine on which I am typing this is an IBM Netvista (8318-31G)
> with an ATI Radeon 7000 PCI graphics card, running Ubuntu 8.04.
> 
> Sometimes it runs for days or even a week without a problem - but it
> also has a tendency to freeze utterly. No response to keyboard or
> mouse, won't accept ssh, not reachable by ping, no response to Alt
> +SysRq. Nothing in /var/log/syslog. Nothing obvious the Xorg log
> either, though the most recent crashed one has this at the end
> 
> Entering Restore TV
> Restore TV PLL
> Restore TVHV
> Restore TV Restarts
> Restore Timing Tables
> Restore TV standard
> Leaving Restore TV
> 
> and the current one doesn't. I can't track this down to any particular
> programme - twice this week it has done it when left at the login
> screen with noone logged on.
> 
> Any bright ideas? The problem seems to have started with 8.04: I never
> had this with 7.10.

I'd say there's a 99% chance that it's hardware. Is the processor fan 
running when it's crashed? Be careful not to shake the machine at all 
while you're taking the cover off to find out - as it may start again 
and you won't know.

Try reseating the ram DIMMs.

It can be extremely difficult troubleshooting seemingly random hardware 
crashes. Good luck!


-- 
http://MaldonIT.co.uk

On 10 Jul, 21:41, Ian  wrote:

> Any bright ideas? The problem seems to have started with 8.04: I never
> had this with 7.10.

Thanks for the suggestions, all. The idea of testing the memory
occurred to me just after I sent the last message, so the machine has
been running the standalone Memtest86+ all night ... no errors. I have
the CPUTemperature applet on my panel, so I'll note what that says
when the next crash happens ... however, the two recent crashes with
no one logged on suggest that it's not load related.

Keep these ideas coming! I'm very grateful.

Ian

Ian wrote:
> On 10 Jul, 21:41, Ian  wrote:
> 
>> Any bright ideas? The problem seems to have started with 8.04: I never
>> had this with 7.10.
> 
> Thanks for the suggestions, all. The idea of testing the memory
> occurred to me just after I sent the last message, so the machine has
> been running the standalone Memtest86+ all night ... no errors. I have
> the CPUTemperature applet on my panel, so I'll note what that says
> when the next crash happens ... however, the two recent crashes with
> no one logged on suggest that it's not load related.

Any PCI/AGP etc cards in it? If so, reseat them. Reseat all motherboard 
connectors too. Give it a good blow out, too - with compressed air (you 
can buy an aerosol can of the stuff if necessary).


-- 
http://MaldonIT.co.uk

On 11 Jul, 07:27, Will Kemp  wrote:

> Any PCI/AGP etc cards in it? If so, reseat them. Reseat all motherboard
> connectors too. Give it a good blow out, too - with compressed air (you
> can buy an aerosol can of the stuff if necessary).

The only card is the graphics one. I shall take your advice!

Ian

Ian wrote:
> The machine on which I am typing this is an IBM Netvista (8318-31G)
> with an ATI Radeon 7000 PCI graphics card, running Ubuntu 8.04.
> 
> Sometimes it runs for days or even a week without a problem - but it
> also has a tendency to freeze utterly. No response to keyboard or
> mouse, won't accept ssh, not reachable by ping, no response to Alt
> +SysRq. Nothing in /var/log/syslog. Nothing obvious the Xorg log
> either, though the most recent crashed one has this at the end
> 
> Entering Restore TV
> Restore TV PLL
> Restore TVHV
> Restore TV Restarts
> Restore Timing Tables
> Restore TV standard
> Leaving Restore TV
> 
> and the current one doesn't. I can't track this down to any particular
> programme - twice this week it has done it when left at the login
> screen with noone logged on.
> 
> Any bright ideas? The problem seems to have started with 8.04: I never
> had this with 7.10.
> 
> Ian

Similar fault here a few years ago and the outcome was an intermittent 
fault on the vid card (Nvidia) changed the card and all is well...



-- 
Regards
  Ted Wager g3tpi
   High Peak UK
    Using Sidux Linux

On 2008-07-11, twager  wrote:

> Similar fault here a few years ago and the outcome was an intermittent 
> fault on the vid card (Nvidia) changed the card and all is well...

I just replaced my fileserver recently, as the original machine
started rebooting at random, even during the boot cycle sometimes.  No
load patterns, sometimes it would reboot and come back up fine and
last for a day or two, other times it would sit there for an hour or
so, constantly rebooting, sometimes reaching linux, sometimes not even
getting that far.  Checked cooling, checked memory, disconnected the
only PCI card (SCSI), in the end I swapped out the motherboard and put
the old memory into the replacement one and it's been fine since,
didn't have any spare processors to swap so it could have been those.

Replaced it in the end with a mac mini I had spare, using 750 gig USB
hard discs!  They're handling sustained disc to disc copies of 20
MBytes/s which is good enough, and if the machine dies all my files
are on discs I can just plug into the next machine in the rack.

-- 
Blast off and strike the evil Bydo empire!
http://youtube.com/user/tarcus69
http://www.flickr.com/photos/tarcus/sets/

Ian wrote:
> On 11 Jul, 07:27, Will Kemp  wrote:
>
>> Any PCI/AGP etc cards in it? If so, reseat them. Reseat all
>> motherboard connectors too. Give it a good blow out, too - with
>> compressed air (you can buy an aerosol can of the stuff if
>> necessary).
>
> The only card is the graphics one. I shall take your advice!

Blowing is bad, the dust can be conductive, sucking is better. If you are a 
smoker, the dust is horrible and more difficult to remove.
Someone I know who's a sixty a day man has always had problems with bad 
physical connections on all his machines. He seems to be continually having 
to reseat them, personally I think there's a connection (or rather there 
isn't ;-) )

On Fri, 11 Jul 2008 09:51:53 +0100, jasee wrote:

> Blowing is bad, the dust can be conductive, sucking is better. If you are a 
> smoker, the dust is horrible and more difficult to remove.
>
Sucking can be destructive too. If your sucker has a
nonconductive plastic nozzle it can generate very high static charges
which can zap your boards, or so I'm told. It sounds plausible, so using a
grounded metal nozzle would make a lot of sense.


-- 
martin@   | Martin Gregorie
gregorie. | 
org       | Zappa fan & glider pilot

Martin Gregorie wrote:
> On Fri, 11 Jul 2008 09:51:53 +0100, jasee wrote:
>
>> Blowing is bad, the dust can be conductive, sucking is better. If
>> you are a smoker, the dust is horrible and more difficult to remove.
>>
> Sucking can be destructive too. If your sucker has a
> nonconductive plastic nozzle it can generate very high static charges
> which can zap your boards, or so I'm told. It sounds plausible, so
> using a grounded metal nozzle would make a lot of sense.

This is of course quite possible with toner as it is designed to hold a 
charge and it can explode (unless you use a special vacuum cleaner with a 
conductive rubber nozzle) and the operator gets shocks, it seems unlikely 
with ordinary dust, though I suppose it's possible.

On 10 Jul, 21:41, Ian  wrote:
> The machine on which I am typing this is an IBM Netvista (8318-31G)
> with an ATI Radeon 7000 PCI graphics card, running Ubuntu 8.04.
>
> Sometimes it runs for days or even a week without a problem - but it
> also has a tendency to freeze utterly.

Thanks again for the suggestions. I decided to have a look inside -
but before I did, I noticed a high pitch scream, right on the edge of
what I can hear. Ah-ha, I thought, that's rather reminiscent of
unseated chips on an Atari ST. [*]

So, top off - and the first thing I spot is that the video car is
sitting at angle, with the pins at one hend barely in the housing.
This did not look good. It's now back where it should be - and I took
the chance to wiggle and reseatch everything else I could see on the
motherboard. Not that that's very much on a SFF Netvista, 'cos it's
all a bit hidden away, but I did a few connectors and the memory.

We shall see ,,,

Ian

On 11 Jul, 15:07, Ian  wrote:

> Thanks again for the suggestions. I decided to have a look inside -
> but before I did, I noticed a high pitch scream, right on the edge of
> what I can hear. Ah-ha, I thought, that's rather reminiscent of
> unseated chips on an Atari ST. [*]

Oops. Forgot this:

* The official (really) cure for which was to drop the computer, flat,
from about six inches onto a hard surface.

And it ran Minix, too. Kids, ask your grandad what Minix was.

Ian

On Thu, 10 Jul 2008 23:48:32 +0000, Tony Houghton wrote:

> See if you can monitor the temperature with lm-sensors and one of its
> front-ends. If not, reboot when everything's warmed up and check it in
> the BIOS.

Munin has a good sensors tracking plugin. You may be able to correlate 
temperative rises and reboots.

-- 
 <http://ale.cx/> (AIM:troffasky) (UnSoEsNpEaTm@ale.cx)
 18:12:16 up 9 days,  4:39,  4 users,  load average: 0.03, 0.03, 0.00
 Convergence, n: The act of using separate DSL circuits for voice and data

On Fri, 11 Jul 2008 17:13:01 +0000, alexd wrote:

> Munin has a good sensors tracking plugin. You may be able to correlate
> temperative rises and reboots.

s/temperative/temperature/

-- 
 <http://ale.cx/> (AIM:troffasky) (UnSoEsNpEaTm@ale.cx)
 18:15:30 up 9 days,  4:42,  4 users,  load average: 0.11, 0.04, 0.01
 Convergence, n: The act of using separate DSL circuits for voice and data

In article ,
	Ian wrote:

> And it ran Minix, too. Kids, ask your grandad what Minix was.

O Tannenbaum, o Tannenbaum,
Wie treu sind dein microkernel!

-- 
Paul Martin

On 11 Jul 2008, Martin Gregorie spake thusly:

> On Fri, 11 Jul 2008 09:51:53 +0100, jasee wrote:
>
>> Blowing is bad, the dust can be conductive, sucking is better. If you are a 
>> smoker, the dust is horrible and more difficult to remove.
>>
> Sucking can be destructive too. If your sucker has a
> nonconductive plastic nozzle it can generate very high static charges
> which can zap your boards, or so I'm told. It sounds plausible, so using a
> grounded metal nozzle would make a lot of sense.

Many years ago a friend of mine who shall remain nameless at a large
biological standards institution took an extremely expensive SGI which
was acting up, took the top off, and applied a sucker.

SNAP rattle gulp! oops. That was a capacitor getting sucked up the tube,
wasn't it?

... cue much frantic soldering and swearing and hoping that the thing
hadn't lost its magic smoke. (It had.)

On 11 Jul 2008, Paul Martin said:
> O Tannenbaum, o Tannenbaum,
> Wie treu sind dein microkernel!

*cough* *choke* Don't *do* that while I'm drinking! I have a very
expensive keyboard here I'll have you know. ;)

Nix wrote:
> On 11 Jul 2008, Martin Gregorie spake thusly:
> 
>> On Fri, 11 Jul 2008 09:51:53 +0100, jasee wrote:
>>
>>> Blowing is bad, the dust can be conductive, sucking is better. If you are a 
>>> smoker, the dust is horrible and more difficult to remove.
>>>
>> Sucking can be destructive too. If your sucker has a
>> nonconductive plastic nozzle it can generate very high static charges
>> which can zap your boards, or so I'm told. It sounds plausible, so using a
>> grounded metal nozzle would make a lot of sense.
> 
> Many years ago a friend of mine who shall remain nameless at a large
> biological standards institution took an extremely expensive SGI which
> was acting up, took the top off, and applied a sucker.
> 
> SNAP rattle gulp! oops. That was a capacitor getting sucked up the tube,
> wasn't it?
> 
> ... cue much frantic soldering and swearing and hoping that the thing
> hadn't lost its magic smoke. (It had.)

Yeah, i'd never risk using a sucker for that sort of reason! Blowing's 
not ideal, but the chances of blowing something off the board are much 
lower than the chances of sucking something off.


-- 
http://MaldonIT.co.uk

In article ,
Nix  <$}xinix{$@esperi.org.uk> wrote:

>Many years ago a friend of mine who shall remain nameless at a large
>biological standards institution took an extremely expensive SGI which
>was acting up, took the top off, and applied a sucker.
>
>SNAP rattle gulp! oops. That was a capacitor getting sucked up the tube,
>wasn't it?
>
>... cue much frantic soldering and swearing and hoping that the thing
>hadn't lost its magic smoke. (It had.)

Well at least he found out why it was (probably) acting up ...

Nick
-- 
Serendipity: http://www.leverton.org/blosxom (last update 6th June 2008)
        "The Internet, a sort of ersatz counterfeit of real life"
                -- Janet Street-Porter, BBC2, 19th March 1996

On 11 Jul 2008, Nick Leverton said:

> In article ,
> Nix  <$}xinix{$@esperi.org.uk> wrote:
>
>>Many years ago a friend of mine who shall remain nameless at a large
>>biological standards institution took an extremely expensive SGI which
>>was acting up, took the top off, and applied a sucker.
>>
>>SNAP rattle gulp! oops. That was a capacitor getting sucked up the tube,
>>wasn't it?
>>
>>... cue much frantic soldering and swearing and hoping that the thing
>>hadn't lost its magic smoke. (It had.)
>
> Well at least he found out why it was (probably) acting up ...

Even if he hadn't, he knew why it was acting up now! :/

On Fri, 11 Jul 2008 13:32:26 +0100, jasee wrote:

> Martin Gregorie wrote:
>> On Fri, 11 Jul 2008 09:51:53 +0100, jasee wrote:
>>
>>> Blowing is bad, the dust can be conductive, sucking is better. If you
>>> are a smoker, the dust is horrible and more difficult to remove.
>>>
>> Sucking can be destructive too. If your sucker has a nonconductive
>> plastic nozzle it can generate very high static charges which can zap
>> your boards, or so I'm told. It sounds plausible, so using a grounded
>> metal nozzle would make a lot of sense.
> 
> This is of course quite possible with toner as it is designed to hold a
> charge and it can explode (unless you use a special vacuum cleaner with
> a conductive rubber nozzle) and the operator gets shocks, it seems
> unlikely with ordinary dust, though I suppose it's possible.
>
From what I heard no toner was needed. In a dry enough atmosphere just
sucking air through an insulating nozzle can generate a charge. Put the
tip near a chip and ZAP! 

Its the same effect as stroking a cat or combing your hair with a plastic
comb and getting sparks or (for those of us old enough to remember nylon
shirts) the static crackle as you pull a nylon shirt off over your head.
 
Back to cleaning computers - sucking or blowing would both generate a
static charge if conditions are right.


-- 
martin@   | Martin Gregorie
gregorie. | 
org       | Zappa fan & glider pilot

Martin Gregorie wrote:
> On Fri, 11 Jul 2008 13:32:26 +0100, jasee wrote:
> 
>> Martin Gregorie wrote:
>>> On Fri, 11 Jul 2008 09:51:53 +0100, jasee wrote:
>>>
>>>> Blowing is bad, the dust can be conductive, sucking is better. If you
>>>> are a smoker, the dust is horrible and more difficult to remove.
>>>>
>>> Sucking can be destructive too. If your sucker has a nonconductive
>>> plastic nozzle it can generate very high static charges which can zap
>>> your boards, or so I'm told. It sounds plausible, so using a grounded
>>> metal nozzle would make a lot of sense.
>> This is of course quite possible with toner as it is designed to hold a
>> charge and it can explode (unless you use a special vacuum cleaner with
>> a conductive rubber nozzle) and the operator gets shocks, it seems
>> unlikely with ordinary dust, though I suppose it's possible.
>>
> From what I heard no toner was needed. In a dry enough atmosphere just
> sucking air through an insulating nozzle can generate a charge. Put the
> tip near a chip and ZAP! 
> 
> Its the same effect as stroking a cat or combing your hair with a plastic
> comb and getting sparks or (for those of us old enough to remember nylon
> shirts) the static crackle as you pull a nylon shirt off over your head.
>  
> Back to cleaning computers - sucking or blowing would both generate a
> static charge if conditions are right.

That's not nearly so much of a worry at this time of year though - as 
there's plenty of moisture in the air to conduct the charge away as it 
builds up. Winter's bad for static - as are air-conditioned environments.



-- 
http://MaldonIT.co.uk

On 11 Jul, 15:07, Ian  wrote:

> So, top off - and the first thing I spot is that the video card is
> sitting at angle, with the pins at one hend barely in the housing.
> This did not look good. It's now back where it should be - and I took
> the chance to wiggle and reseatch everything else I could see on the
> motherboard. Not that that's very much on a SFF Netvista, 'cos it's
> all a bit hidden away, but I did a few connectors and the memory.
>
> We shall see ,,,

About to reboot for a kernel update, but ...

ian@squirrel:~$ uptime
10:04:24 up 4 days, 23:23,  3 users,  load average: 0.64, 0.43, 0.19

Many thanks for the advice, folks.

Ian