Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site brl-tgr.ARPA
Path: utzoo!linus!philabs!cmcl2!seismo!brl-tgr!internet!stanonik@nprdc
From: stanonik@nprdc
Newsgroups: net.unix-wizards
Subject: ethernet puzzle
Message-ID: <5832@brl-tgr.ARPA>
Date: Fri, 16-Nov-84 11:20:35 EST
Article-I.D.: brl-tgr.5832
Posted: Fri Nov 16 11:20:35 1984
Date-Received: Sat, 17-Nov-84 20:06:17 EST
Sender: news@brl-tgr.ARPA
Organization: Ballistic Research Lab
Lines: 30

Ok puzzle fans, we're having problems etherneting between vaxs and
suns, all running 4.2bsd.  Our localnet looks like:

pacific nprdc                       suna sunb sunc sund
   |      |        ~105 meters        |    |    |    |
-------------------------------------------------------

All transceivers are 3com, and both vaxs have interlan interfaces.
1) nprdc can talk to suna and sunc, but not sunb and sund.
2) pacific can talk to all of the suns.
3) if we swap sun transceivers (eg, suna and sunb), the problem seems 
to follow the transceiver (eg, nprdc and sunb can talk, but nprdc and
suna can't).
4) if we swap vax transceivers, the problem seems to follow nprdc.
5) the vaxs can talk to one another.
6) the suns can talk to one another.
7) by "can't talk", I mean that even arp packets are not received.
8) netstat -i shows no collisions on any of the machines.

3 seems to imply a problem with the sun tranceivers, but 4 seems to
imply a problem with nprdc's interlan interface.  On the other hand,
we don't understand how an interface could talk to some transceivers, 
but not others.  Before we send nprdc's interface in for repair (which
means pulling nprdc off the localnet), we'd like to convince ourselves 
that it is the problem.  Any similar experiences?  Any suggestions?

Thanks,

Ron Stanonik
stanonik@nprdc