Re: Linux Sudden Death Syndrome?

Matthew J Brown (mjb@sophos.com)
Wed, 11 Oct 1995 10:09:24 +0100

Matthias Sattler writes:
> Hmmmm... after reading this discussion for a while, I think that there
> really might be a serious bug in the linux-kernel somewhere. It is totally
> possible that is shows up only in a certain environment, what is the only
> constant thing in the tests that were made (What services do you use?
> What other machines are connected to the network? Do you use non standard
> (eg. self written) programs?.....).

Well, in the last week I've had two dead hangs on my home machine
under 1.3.30 ... but in the same period, my work machine has been up
continuously under much heavier use with no problems.

Of course, this could be an X hang, not a Linux one ... I think it is
possible for X to hang in such a way as you can't get back to another
VC. Without a serial terminal, of course, it's a bit difficult to
test this. (tho not impossible of course)

Which brings me to another thing -- do those suggestions by Alan Cox
about how to get kernel debugging info on a crash to print on screen
work if you're under X when the crash happens?

[the home machine is 486/66, 8mb ram, soundblaster 16, ide cdrom, SLIP
but no PPP; both hangs happened while dialed up to the Net. Work
machine doesn't have idecdrom, sb16 or SLIP but does have 3c509 and
future domain 16xx scsi with a 1gb disk off it, and 20Mb memory.]

-Matt