Path: utzoo!bnr-vpa!bnr-fos!bmers58!schow
From: schow@bmers58.UUCP (Stanley Chow)
Newsgroups: news.admin
Subject: Re: Noise in Message ID's causing dupes
Summary: Is there not a checksum??
Message-ID: <284@bmers58.UUCP>
Date: 29 Sep 89 19:30:36 GMT
References: <14751@bfmny0.UU.NET>
Reply-To: schow%BNR.CA.bitnet@relay.cs.net (Stanley Chow)
Distribution: na
Organization: Bell-Northern Research, Ottawa, Canada
Lines: 56
Followup-To:
Keywords:

In article <14751@bfmny0.UU.NET> tneff@bfmny0.UU.NET (Tom Neff) writes:
>OK, here's an example.  Someone trashed a message ID en route,
>enabling B news to post a dupe.
> [extra info deleted.]
>! Path: bfmny0!uunet!ginosko!gem.mps.ohio-state.edu!tut.cis.ohio-state.edu!ucbvax!ARISTOTLE-GW.JPL.NASA.GOV!pjs
>! Message-ID: <8909271818.AA01035@aristotle.Jpl.Nasa.Gov>
>              
>! Path: bfmny0!uunet!wuarchive!brutus.cs.uiuc.edu!apple!bionet!ames!lll-lcc!unisoft!mtxinu!ucbvax!ARISTOTLE-GW.JPL.NASA.GOV!pjs
>! Message-ID: <890=271818.AA01035@aristotle.Jpl.Nasa.Gov>


First of all, I am a relatively new *user* of Usenet and don't know a lot
about the mysteries of the various news software and the protocols. Since
this business of the duplicate articles is getting annoying and many people
have looked at it with no resolution, I figured throwing in my guesses 
won't hurt anything.

I am a little surprised this can happen. Isn't there a checksum somewhere
that covers the message ID? Assuming (some news software) does not have 
checksum protecting the message ID (or that some idiot disabled it), then
I have a scenerio that may explain the problem.

The history of the article have to be something like:

	       -- Site3, Id2 --
              /                \
--> Site1, Id1                 Site4, Id1+Id2
	      \                /
	       -- Site2, Id1 --

The error would have happened on the transmission from Site1 to Site3. 
(This assumes each site only transmites one copy of any article to its
downstream sites).

Following this reasoning, the duplicate article in question must have
originated during the transmission from ucbvax to mtxinu. All other sites
are propagating articles as they should. 

It is possible (barely) for the error to have been JPL -> ucbvax, but this
requires that JPL thinks the first transmission failed but ucbvax to have
accepted the first article.

One confirmation  would be to look at the path of the erroneous article
at the "innocent" sites. The path should be "abnormal" in that it took
a longer path of the form  originator!..!Site1!Site3!...!Site4!..Site2.

Another confirmation would be to look at the log files of the suspected
link. It should have lots of transmission errors. (Assuming that at least
the bodies of articles are protected by checksum).


-- 
Stanley Chow        BitNet:  schow@BNR.CA
BNR		    UUCP:    ..!psuvax1!BNR.CA.bitnet!schow
(613) 763-2831		     ..!utgpu!bnr-vpa!bnr-fos!schow%bmers58
Me? Represent other people? Don't make them laugh so hard.