chris@umcp-cs.UUCP (10/16/83)
Somewhere along the path:
umcp-cs!cvl!rlgvax!seismo!philabs!cmcl2!floyd!clyde!ihnp4!hlexa!hlhop!
mhuxh!mhuxi!houxm!hogpc!houti!ariel!vax135!allegra!linus!decvax!duke!unc!
achilles!ulysses!cbosgd!mark
one or more machines delayed an article posted on September 26 until now,
October 16. With so many sites in that path, it's probably going to
be very hard to track down. I know that it wasn't at seismo, rlgvax,
or cvl, or the sending site, cbosgd. Can anyone tell where the article
got stuck for two weeks?
*** update *** (edited in by hand) Looks like the problem is one of
floyd, clyde, or cmcl2 [maybe, I'm not positive]. Anyone from those
sites care to comment?
--
In-Real-Life: Chris Torek, Univ of MD Comp Sci
UUCP: {seismo,allegra,brl-bmd}!umcp-cs!chris
CSNet: chris@umcp-cs ARPA: chris.umcp-cs@UDel-Relayevans@wivax.UUCP (Barry Evans) (10/17/83)
Earlier, Chris Torek said that there were old (duplicate) articles coming from
somewhere along the path:
umcp-cs!cvl!rlgvax!seismo!philabs!cmcl2!floyd!clyde!ihnp4!hlexa!hlhop!
mhuxh!mhuxi!houxm!hogpc!houti!ariel!vax135!allegra!linus!decvax!duke!unc!
achilles!ulysses!cbosgd!mark
He also narrowed it down to:
floyd, clyde, or cmcl2 [maybe]
Well here at wivax we have, and still are, receiving many *old* articles since
Sept. 23 or there abouts. This all began Sunday (Oct 16). I did some
checking of the paths of many of these diseased articles, and narrowed it
down to either floyd or cmcl2, but probably not cmcl2. (I say probably not
cmcl2 only because they were not found in 'one` of the paths - but there could
exist human/machine error in there).
I did narrow out clyde and philabs which were also high on the list, probably
because of their close affiliation with floyd and cmcl2.
Now, I'm not saying this is definately floyd's (or cmcl2's) fault, I'm just
saying that they seem to be the source, as seen from wivax. Anybody there
or elsewhere able to confirm/deny my findings?
-barry
--
Barry Evans {decvax,linus}!wivax!evans
Wang Institute (617) 649-9731 x383pcl@ihnp4.UUCP (Paul C. Lustgarten) (10/18/83)
I can explain a *portion* of the duplicate articles that have hit the
net lately - specifically, those in net.news.group that have the
transition ...!hlexa!ihnp4!... in their path. (This is not a normal
link for through traffic.) The situation was that ihnp4 had not been
storing articles in net.news.group due to a missing active file entry.
(We understood that it was net.news.groups.) When I discovered & fixed
this, I shipped ihnp4 the existing articles from that newsgroup on
hlexa. Unfortunately, I forgot that we don't expire 'net.news' on
hlexa, and thus some of the articles were a bit more than two weeks old
(luckily not many, because I had culled out the really old stuff some
weeks before).
I don't think this was a significant number of duplicates, but decided
it might help those trying to track down the *other* duplicates that
seem to have appeared at about the same time, to know that there are at
least *two* sources of old articles involved here.
Sorry for any undue puzzlement this may have caused.
Paul Lustgarten
Netnews Administrator for
AT&T Bell Laboratories, Short Hills NJ (the 'hl*' sites)
{hlexa, ihnp4} !pcl