Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!uunet!comcon!tim
From: tim@comcon.UUCP (Tim Brown)
Newsgroups: comp.unix.i386
Subject: TCP  Clarification
Keywords: tcp nfs
Message-ID: <135@comcon.UUCP>
Date: 5 Dec 89 04:51:34 GMT
Organization: Computer Connection
Lines: 30

I asked for help with a tcp problem, I would like to clarify.

First I am running Interactive 2.0.2, thier host based tcp and nfs.

The problem is that one of the machines reboots during some as yet
undetermined activity.

machine one loks like this:
	386/20 MHZ AMI based clone with 8MB ram, two 122MB Harddisks
machine two (the one that re-boots)
	386/33MHZ Micronics based clone with 12MB, one 330MB harddisk.

I can make the second machine reboot by starting some heavy activity
between them, such as this command from machine one:
	find some_dir_on_machine_two -print|cpio -o >/dev/null
That gets them going, but does not cause the reboot, if while that is
happening I rlogin to machine two from machine one bingo!  Machine two
will start a reboot just like I have reached over and hit the reset
switch!

A little background:
	-	I have pushed all of the tunables really high.
	-	I checked netstat -m output to verify that the tunables
		are staying in range.
	-	Streams tunables are adequately defined.
Anyone have an idea?

Tim Brown                           |
Computer Connection                 |
(attmail or uunet)!comcon!tim       |