Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!mcsun!ukc!keele!csd35
From: csd35@seq1.keele.ac.uk (Jonathan Knight)
Newsgroups: comp.protocols.tcp-ip
Subject: Why do I need multiple host names for the same host?
Message-ID: <705@keele.keele.ac.uk>
Date: 8 Nov 90 12:35:53 GMT
Organization: University of Keele, England
Lines: 37

Our dept has just re-arranged its ethernet into 2 seperate ethernets
with a sun 4 acting as the router.  It is the only host which has
more than one cable attached to it.

Following the manual we have put:
192.42.100.3    do loghost
192.82.242.1    doc_1

into our hosts table.  This causes problems as some of our diskless
stations need to be told to get their swap and root from 'doc'
and others need to get it from 'doc_1'.  This is caused by the fact
that they have no routing information at the point at which they want
to download vmunix.  Therefore the host they boot from must be on
their network.

I have two questions.

1)	When a host on the 192.82.242 network tries to talk to
	'doc' is the packet addressed to 192.42.100.3?  If not
	is this packet unpacked by doc?  If not does this packet
	actually get placed on the 192.42.100 network and re-read
	by doc?

2)	Is there a way to indicate to each host that doc is local
	on their ethernet?

The only solution I came up with was having multiple /etc/hosts, one
for each network with doc listed on that network.  This is inelegant as
we want to use YP (sorry - NIS) to distribute the known hosts to all
networks and we want to avoid multiple domains.

Any help will be gratefully received.
-- 
  ______    JANET :jonathan@uk.ac.keele.cs     Jonathan Knight,
    /       BITNET:jonathan%cs.kl.ac.uk@ukacrl Department of Computer Science
   / _   __ other :jonathan@cs.keele.ac.uk     University of Keele, Keele,
(_/ (_) / / UUCP  :...!ukc!kl-cs!jonathan      Staffordshire.  ST5 5BG.  U.K.