Re: What are the most common causes of linux system hangs?
Hello Joe,
On Fri, 14 Apr 2000, Joe Emenaker wrote:
> Currently, I'm having this problem with one. Just... out of the blue, it
> will hang dead in its tracks. The keyboard doesn't even wake the screen so I
> can see if there are any kernel panic messages or anything. Ctrl-Alt-Del
> doesn't do anything. I have to hit the rest button.
>
> Now, I'm pretty certain that it's some hardware problem. But, I'd like to
> avoid moving the whole system to a brand-new machine, find that the problem
> has gone away, and conclude that there's just *something* bad about the old
> server and that I need to chuck the whole thing.
>
> So, I'd like to isolate the problem, if I could.
>
> With that in mind, does anyone have any personal experience concerning what
> the problem usually is in these cases? Motherboard? RAM? Has it ever helped
> anyone to *under*clock the CPU?
>
> I'm anxious for any ideas....
I can only tell you about some experiences I made with an old 486,
which I got very cheap without a harddisc. I bought an IDE-HD and put it
in, installed debian and I also got occasional system crashes with exactly
the same symptoms as you reported, except that it always occured, when
there was heavy I/O on the harddisc, and afterwards the hd-LED was always
on. I also tried different settings, with the demons and so on, but it
didn't help. As I had a small DOS-Partition on the hard-drive, I booted
into DOS, just to test. The problem still occured. So it was quite clear
to me, that it should be a hardware matter.
In my case, it was easy to isolate the problem. I assumed, that the
computer had worked fine in the office, it had been used before (as some
kind of diskless terminal, booting from a network). So it was clear to me,
that it had to do something with the harddisc which I put in. First
thing, I did, was change the connector cable. Didn't help. As the harddisc
was new, I supposed, it could be a problem with the I/O-card, which was in
the computer. So I exchanged that one, and guess what? The thing worked
fine.
So in essence, I agree with you, that your Problem might most probably be
a hardware matter. But it could be very difficult to isolate the problem.
I am not an expert in hardware matters, but I don't think that
underclocking would be the way to go. If your motherboard and CPU fit
together, and they support the clock you are using, this shouldn't be a
problem (There was another mail on the debian-user list just recently,
saying that W95 was falling over on AMDs faster than 350MHz on certain
motherboards, but we're not talking about windoze in your case, and I
don't think this is a very common case).
I don't know, how BIOS RAM-Tests work, but I would first of all conclude,
that if this one never reports a problem, it should not be a problem with
the RAM.
Don't count on my tips too much, as I am not a techie, but just a "normal
user". But I just wanted to show you, how I would try to isolate the
problem:
Try removing or exchanging (one by one) those things from your system,
that could cause the crashes. Start with the components, which are most
likely to cause the problem, and which can most easily be
removed/exchanged (For example easy to exchange: connector cables). This
can be a very time consuming thing, but I don't see another way of
isolating hardware problems of such a kind, when you don't get error
reports or kernel-messages in case of a crash.
So this mail is perhaps not of much help for you, as I can't say: "This or
that is definitely the thing to look at". But I also think, that this
would be impossible to do, as it can be many things in the inside of your
computer, which can cause such problems.
Regards,
Daniel
Reply to: