IP address conflict - SUMMARY

From: Carlo Tiana (carlo@white.stanford.edu)
Date: Wed Feb 06 1991 - 13:37:14 CST


Thanks to all those (many) who replied. The original message was:

>Subject: duplicate IP address

>Hello all!
>We have a problem recurring at random times that has been causing
>one of our 4/110's to hang for up to a few hours at a time. The
>console message is:
>
>Feb 5 13:03:50 helmholtz vmunix: duplicate IP address!!
>sent from ethernet address: 2:7:1:4:1c:26
>
>helmholtz is - of course - our machine, while the ethernet address
>shown is unknown to us. The obvious suspicion is that some machine
>somewhere with that ethernet address is using our internet address
>and confusing the hell out of our machine. Not being at all an arp
>or rarp guru I have not been very successful at finding out more
>information about the offending machine, since our tables contain
>the name of our machine as corresponding to our internet address.
>I am working on the assumption that knowing the machine name would
>be a step forward in attempting to locate it, but basically I am
>really clutching for straws here. Any help would be appreciated.
>Thank you very much,
>Carlo.
>
>carlo@white.stanford.edu
>carlo@vision.arc.nasa.gov
>
>PS - I am also told that from the ethernet number we can at least
> figure out the machine vendor. Is this true? Anyone know how?
>PPS - If it's *your* machine doing this, *stop it*! :-)

The replies suggested/commented one or more of the following:
-shutdown/interrupt the "victim" machine and try to telnet to the "culprit"
 machine: the login banner may offer more info as to its whereabouts or
 ownership. (We had tried this but were never successful - read on).
-hangup (SIGHUP) to inetd (no chance - the "victim" machine was *dead*; I
 mean no remote logins, not even mouse updates or screen refreshes).
-shutdown/interrupt the "victim" machine and do "arp -a" or use etherfind
 on some other machine to fine what info was stored in the tables.
-a lot of replies told me that the problem must arise on my segment of
 Ethernet, and not from some place far away (this was a big hint).
-a lot of replies told me what the board maker was likely to be - there are
 2 or 3 possible makers, but many pointed out Racal/Interlan. This is what
 clinched it.
I had noticed that the problem arose at least most of the time when I was
under pressure to do something on the the "victim" machine, and not at
other times. I originally put that down to Murphy's law, but alas, the
person who was putting on the pressure is the owner of a Mac with a Racal
Interlan Ethernet board inside it!! So as he was trying to do something and
asking me to do something else at my end, *that's* when the problems
occured. Moral of the story: the "victim" machine was down at the time the
Mac's board was purchased, so the sys admin "borrowed" the IP address for
that board (and never bothered to get another, permanent one assigned). Of
course the Sun eventually came up again, and the rest is history: at the
rare times when the Mac's board is used, the problem occurs. So much for
people who supposedly know what thay are doing.
In general, people pointed out that this is a very difficult situation to
rectify.

One other piece of information was that the list of Ethernet borad makers
with their associated Ethernet number triplet is available for ftp from:
mic.ddn.mil; cd to rfc:, and get assigned-numbers.txt. Thank you, Doug Kratky
<kratky@electra.boeing.com>.

As usual, this list proved very useful and very timely so too: as usual, I
started getting replies before my own posting arrived back to me!

Credits:

stpeters@dawn.crd.ge.com (Dick St.Peters)
dupuy@hudson.cs.columbia.edu (Alexander Dupuy)
Rich Wales <wales@CS.UCLA.EDU>
geertj@ica.philips.nl (Geert Jan de Groot)
tom@yac.llnl.gov tom slezak
Ken Rossman <ken@watsun.cc.columbia.edu>
del@mlb.semi.harris.com (Don Lewis)
Charles <mcgrew@aramis.rutgers.edu>
dlc@c3.c3.LANL.GOV (Dale Carstensen)
fwp1@CC.MsState.Edu
bala@Synopsys.COM (Bala Vasireddi)
eap@bu-pub.bu.edu (Eric A Pearce)
mrwallen@UCSD.EDU (Mark R. Wallen)
smb@ulysses.att.com
mmikulska@UCSD.EDU (******actually got the same message 4 times!?******)
kevin@Corp.Sun.COM (Kevin Sheehan {Consulting Poster Child})
rich%idacrd@Princeton.EDU (Richard Schultz)
oconnor@sccgate.scc.com (Mike O'Connor)
torda@igc.ethz.ch (Andrew Torda )
paula@atc.boeing.com
randy@ncbi.nlm.nih.gov (Rand S. Huntzinger)
curt@ecn.purdue.edu (Curt Freeland)
selig@centaur.msfc.nasa.gov (Bill Selig - SysAdmin)
Steve Hanson <hanson@pogo.fnal.gov>
trinkle@cs.purdue.edu
Doug Kratky <kratky@electra.boeing.com>
blc@sol.med.ge.com (Brett Chapman x5156)
dlc@c3.c3.LANL.GOV
bob@MorningStar.Com
dan@BBN.COM
stern@East.Sun.COM (Hal Stern - Consultant)
sparc2!ts@sacto.West.Sun.COM (Troy Schumaker)



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:10 CDT