strange freeze

strange freeze

[Elias Athanasopoulos]

Hello!

I know this is a little bit off topic, but I also know that the list is
being followed by some very skillfull admins, whose experience might help. :-)

I have a dual box with Windows XP and SuSE Linux 8.2. Each operating system
has its own hard disk. Both of the hard disks are IDE.

The problem is that the box suddenly freezes. If it is on Windows it 
freezes (i.e. no mouse movement, no disk activity, num-lock can't change
state, i.e. the keyboard is dead [*]) and after 10-15 secs it gets back to
normal. On Linux it freezes (exactly the same symptoms) forever.

My guess is that it's a hardware or a power supply related problem. I 
exclude the last case, because noone else in the office has faced such a
problem and I even changed the cable.

So, I have to find if it has to do with the CPU, memory, hard disk or even
the graphics card's memory. As far as the hard disk case is concerned, I have
noticed that Windows get frozen, randomly, but *mostly* when a file request is 
opened. OTOH, on Linux I can reproduce the problem when I run the updatedb program. 
But, each OS is in a separate disk! Of course during the updatedb process the
Windows fs (FAT32 not NTFS) is beeing read (it is mounted at boot time), but I 
hardly can believe a damaged fs can freeze the whole system.

ANW, if anyone has anything to suggest I'll be happy to hear him/her. I'm looking
for a way to find the root of a problem by using some Linux diagnostic tools (I'll
start with memcheck86), because I know quite a little about Windows. So, suggestions
for hardware diagnosis via Linux software is welcomed, too. :-)

Thank you for your time!

Regards,
Elias

[*] The CPU cooler works.

-- 
University of Athens			I bet the human brain 
Physics Department				is a kludge --Marvin Minsky 

	
-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs

Re: strange freeze

[Ray Olszewski]

If your problem arises with both Windows and Linux, it is almost surely a 
hardware problem. But your message lacks any detail about the hardware 
involved, leaving us only with conjecture.

I have seen something at least superficially similar here, and it had to do 
with running large hard disks without enabling DMA. On the Linux side, this 
can happen if the mobo has an IDE chipset that the kernel you are using 
does not support, or if the kernel is not compiled to use DMA by default.

Again on the Linux side, if this is the cause, the failures can appear to 
be random, because they occur when the kernel has to do a large sync to 
disk (or use swap), something it does its best to do invisibly. updatedb 
produces heavy disk access, which is why it causes the problem reliably.

To investigate this guess from the Linux side, use

         hdparm /dev/hda

(replacing hda with the right designators for your drives) to see if DMA is 
active. If it is not, see if you can turn it on with

         hdparm -d 1 /dev/hda

(same parenthetical applies) and, if you can, if it fixes the problem.

Other things to try ...

You might also try provoking the problem while you have "top" running on a 
display that will survive the hang (e.g., an ssh session from another host) 
to see what the system thinks it is doing when it hangs.

You might explore how thoroughly hung the system is ... for example, does 
the kernel still respond to pings (something that will survive often if 
userspace has crashed beyond hope)?

If you physically disconnect the Windows drive, does the Linux system still 
exhibit this problem? And vice versa?

If that is not it ... well, analysis requires information, so perhaps your 
telling us the details of the hardware, and on the Linux side, the 
partitioning setup and fstab entries,  would suggest something. Be specific 
about cabling too ... does the BIOS see the drives as UDMA, for example? 
Are you using the right kind of cable, and are you relying on cable select 
or setting drives as master and slave with jumpers? (I'm assuming the 
drives share a DMA channel because you said you changed "the cable", not 
"the cables".)

Finally ... "forever" is a meaningless term in a trouble report; how long 
do you actually wait before rebooting? When I had my DMA problem I saw 
waits of several minutes at times before I got a prompt.

At 09:11 PM 5/7/2003 +0300, Elias Athanasopoulos wrote:
>Hello!
>
>I know this is a little bit off topic, but I also know that the list is
>being followed by some very skillfull admins, whose experience might help. :-)
>
>I have a dual box with Windows XP and SuSE Linux 8.2. Each operating system
>has its own hard disk. Both of the hard disks are IDE.
>
>The problem is that the box suddenly freezes. If it is on Windows it
>freezes (i.e. no mouse movement, no disk activity, num-lock can't change
>state, i.e. the keyboard is dead [*]) and after 10-15 secs it gets back to
>normal. On Linux it freezes (exactly the same symptoms) forever.
>
>My guess is that it's a hardware or a power supply related problem. I
>exclude the last case, because noone else in the office has faced such a
>problem and I even changed the cable.
>
>So, I have to find if it has to do with the CPU, memory, hard disk or even
>the graphics card's memory. As far as the hard disk case is concerned, I have
>noticed that Windows get frozen, randomly, but *mostly* when a file 
>request is
>opened. OTOH, on Linux I can reproduce the problem when I run the updatedb 
>program.
>But, each OS is in a separate disk! Of course during the updatedb process the
>Windows fs (FAT32 not NTFS) is beeing read (it is mounted at boot time), 
>but I
>hardly can believe a damaged fs can freeze the whole system.
>
>ANW, if anyone has anything to suggest I'll be happy to hear him/her. I'm 
>looking
>for a way to find the root of a problem by using some Linux diagnostic 
>tools (I'll
>start with memcheck86), because I know quite a little about Windows. So, 
>suggestions
>for hardware diagnosis via Linux software is welcomed, too. :-)



-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs

Re: strange freeze

[Elias Athanasopoulos]

Hello Jos, Ray!

On Wed, May 07, 2003 at 09:42:08PM +0200, Jos Lemmerling wrote:
> When did you experience the first time these problems? Maybe after you
> attached a second hard-drive? Then it's probarly a power-supply
> issue... Try a heavier powersupply (more watts).

Exactly.

> A nice test would be to disconnect the windows-hdd and try if linux
> freezes (and vise versa).

I did it. That was it. :-) Now, I have to convince the technical department
of my office to replace my power supply unit with a better one. This is *less*
trivial than finding the actual problem. :-)) 

Thank you both for the great feedback. Too bad, IIRC, that linux-newbie has
no archives. Lots of e-mail-responses have great value. Maybe we should 
convince someone to donate some web space...

Regards,
Elias

-- 
University of Athens			I bet the human brain 
Physics Department				is a kludge --Marvin Minsky 

	
-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs

Re: strange freeze

[Eric]

How can you exclude the power supply just becuase no-one else in the office 
has the same problem? Unless I mis-understood you the power supply should 
still be held in suspect. If you can get away with it, swap out the PS with 
someone elses in the office or bring one from home. Is it a terribly 
low-wattage supply 150-200? Try unpluging one of the disks when botting to 
the other OS, its not much of a power drai, but its still something.

-- 
#----------------------
#Eric Bambach   
#Eric@CISU.net 
#----------------------
-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs

Re: strange freeze

[Mike Dresser]

On Tue, 13 May 2003, Eric wrote:

> How can you exclude the power supply just becuase no-one else in the office
> has the same problem? Unless I mis-understood you the power supply should
> still be held in suspect. If you can get away with it, swap out the PS with
> someone elses in the office or bring one from home. Is it a terribly
> low-wattage supply 150-200? Try unpluging one of the disks when botting to
> the other OS, its not much of a power drai, but its still something.

Indeed, I've got a system at the office running W2k that would freeze, p4
2.4.

I even brought over Linux on a spare hard drive, and booted off that,
still locked up.(ruling out software)

Replaced all the hardware with spares, and tried again.  Machine still
locked up same as before.

Turns out that it looks like the battery backup is likely defective,
causing voltage spikes/drops that are locking up the system.

Next time I grab the Fluke and leave that plugged into the output of
the UPS overnight!

Mike

-
To unsubscribe from this list: send the line "unsubscribe linux-newbie" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.linux-learn.org/faqs