> On Tue, 23 Feb 1999, Kevin M. Lahey wrote:
> 
> > I'm still not *totally* convinced that it isn't hardware;  these are
> > moderately funky SuperMicro Pentium Pro motherboards with I2O and 256MB 
> > of RAM.  I'll try to reproduce this on some other hardware.
> > 
> > Just another "Me, too",
> 
> Yet another "me too". Using ab and apache, using localhost w/o keepalives,
> 10000 transfers with 10 clients, my ASUS Super socket 7 board spontaneously
> reboots as well. The only clue that impending doom was about to occur is
> the occasional "out of socket buffers" message that ab returns when it
> fails. If I then restart the ab, it reboots, bypassing ddb. netstat -m has
> not shown anything out of the ordinary. I haven't filed a bug report yet
> because I've been hoping that you[kml] or I might find something more
> concrete to report. 
> 
> In short, I don't think that it's hardware.

	I considered hardware, but if other people are seeing it on different
hardware, I don't believe that's what it is.  Anyone who wants to test this can
grab a copy of websplat from http://bleu.west.spy.net/~dustin/soft/ and try it
out.  If it doesn't work, modify sockets.c to not disable the nagle algorithm
(I'm going to make this a runtime option, just haven't yet).  I did not disable
the nagle algorithm when I was doing my tests.

--
SA, beyond.com           My girlfriend asked me which one I like better.
pub  1024/3CAE01D5 1994/11/03 Dustin Sallings <dustin@spy.net>
|    Key fingerprint =  87 02 57 08 02 D0 DA D6  C8 0F 3E 65 51 98 D8 BE 
L_______________________ I hope the answer won't upset her. ____________