Re: Linux Sudden Death Syndrome?

Hendrik Visage (Hendrik.Visage@VECTOR.CO.ZA)
Thu, 5 Oct 1995 13:46:53 +0200 (SAT)

> Hendrik Visage said:
> >
[Discription deleted]
> I think the problem that everyone is seeing is from lack of memory and the
> linux kernel malloc(). You can hang any machine, any hardware simple by
> booting with the option mem=4M, and then compiling something big like wine
> over nfs with 8k rsize and wsize.

I'm not quite satisfied with the memory shortage suggestion:

NE2000 Clone
ET4000 Clone

+- 5-8MB buffers according to Top! (On the rlogin session to the machine at LSDS
{Linux Sudden Death Syndrome})

Running processes:
ftp or cp over NFS mounted dir.
Filesize >>20MB.

The malloc bug stuf I might believe, cause it will make a bit of sense if you
say that (In my case the network giving trouble with bad TX/RX packets) the
reallocation of TCP/IP buffers for the bad or out of order packets, might make
the bug trigger.

BTW: those that DID found the LSDS, was that with a.out/ELF, which gcc compiled
kernel, which libc, Memory size, Mem usage, What type of processes where

Reply to me with Subject line: LSDS
and I'll see if there might be any correlation

Groetend / Sincerely Yours

Hendrik Visage
#include <Standard/Disclaimer>
Vector Customer Support
+27 11 315 4330