wcf@psuhcx (Bill Fenner) (08/03/88)
Is it really necessary to list 1,700 internal at&t sites in att's map
entry? I figured I should check for paths going through att so nobody's
mail gets discarded, and I came up with over 2,000... so I added -datt to
the pathalias command line, and still came up with over 1,700... looked
at them all and they all (seemed to be) at&t internal sites. Now, can't
you just say
att .att.com
and then let internal sites be addressed as xxx.att.com? It would make
it a lot easier to check for paths that are going through att that shouldn't
be, and would make the paths file all that much smaller.
Just my 2 cents..
Bill
P.S. sorry for posting this in news.admin, I know it has nothing to do with
MES... :-)lyndon@ncc.Nexus.CA (Lyndon Nerenberg) (08/04/88)
In article <313@psuhcx.psu.edu> wcf@psuhcx (Bill Fenner) writes: >Now, can't you just say >att .att.com >and then let internal sites be addressed as xxx.att.com? It would make >it a lot easier to check for paths that are going through att that shouldn't >be, and would make the paths file all that much smaller. The trouble is the System V mailers don't grok domain addresses. If AT&T was to do this, they would have to run software from the GNU project :-) -- VE6BBM {alberta,pyramid,uunet}!ncc!lyndon lyndon@Nexus.CA
wcf@psuhcx.psu.edu (Bill Fenner) (08/05/88)
In article <10368@ncc.Nexus.CA> lyndon@ncc.nexus.ca (Lyndon Nerenberg) writes: |In article <313@psuhcx.psu.edu> wcf@psuhcx (Bill Fenner) writes: |>Now, can't you just say |>att .att.com |>and then let internal sites be addressed as xxx.att.com? It would make |>it a lot easier to check for paths that are going through att that shouldn't |>be, and would make the paths file all that much smaller. | |The trouble is the System V mailers don't grok domain addresses. If AT&T |was to do this, they would have to run software from the GNU project :-) | If att is running sendmail (I suppose that's a silly thought?) it wouldn't be too hard to create another mailer, say 'attuucp', which can strip all of the .att.com 's out of the headers to hand to the internal site, and then add them back when going out... Bill
bob@allosaur.cis.ohio-state.edu (Bob Sutterfield) (08/06/88)
In article <313@psuhcx.psu.edu> wcf@psuhcx (Bill Fenner) writes: >Now, can't you just say >att .att.com >and then let internal sites be addressed as xxx.att.com? Since we've got a DIRECT connection across town to an att, osu-cis's `paths' file already says .att.com att!%s They handle everything internal to ATT for us. In article <10368@ncc.Nexus.CA> lyndon@ncc.nexus.ca (Lyndon Nerenberg) writes: >The trouble is the System V mailers don't grok domain addresses. If AT&T >was to do this, they would have to run software from the GNU project :-) Not hardly. Not that your suggestion would be all bad, but at least some ATT sites already do domains correctly: Received: by att.ATT.COM (smail2.5 - att-cb) Others, though, require users to do something like arpa!host.dom.ain!user to get out into the internetworking world. It's a big organization - please don't consider it monolithic. -=- Bob Sutterfield, Department of Computer and Information Science The Ohio State University; 2036 Neil Ave. Columbus OH USA 43210-1277 bob@cis.ohio-state.edu or ...!{att,pyramid,killer}!cis.ohio-state.edu!bob
wcf@psuhcx.psu.edu (William C. Fenner) (08/08/88)
In article <19426@tut.cis.ohio-state.edu> bob@allosaur.cis.ohio-state.edu (Bob Sutterfield) writes: |In article <313@psuhcx.psu.edu> wcf@psuhcx (Bill Fenner) writes: |>Now, can't you just say |>att .att.com |>and then let internal sites be addressed as xxx.att.com? | |Since we've got a DIRECT connection across town to an att, osu-cis's |`paths' file already says | .att.com att!%s So does ours. Unfortunately, it also seems to have over a thousand individual internal att sites with the path psuvax1!aecom!att!site. What I'm complaining about is these thousand sites. Bill
wes@obie.UUCP (Barnacle Wes) (08/08/88)
In article <313@psuhcx.psu.edu> wcf@psuhcx (Bill Fenner) writes: % Now, can't you just say % att .att.com % and then let internal sites be addressed as xxx.att.com? It would make % it a lot easier to check for paths that are going through att that shouldn't % be, and would make the paths file all that much smaller. In article <10368@ncc.Nexus.CA>, lyndon@ncc.Nexus.CA (Lyndon Nerenberg) replies: > The trouble is the System V mailers don't grok domain addresses. If AT&T > was to do this, they would have to run software from the GNU project :-) Right, but `smail' does understand domain names. Anybody who has a purely UUCP System V installation and doesn't use smail is lacking some smarts. It is very easy to get - it's distributed in comp.sources.unix fairly often. Smail is one of the few PD unix tools I've found that solves more problems than it creates. Now if we could just get the map to be a little more realistic. While we're on the subject, is site tut.cis.ohio.edu passing ANY mail to ANY other sites at all? Everything I've sent through them lately has bounced, to the point where I re-ran pathalias with tut marked as 'dead.' Wes Peters System baby-sitter for `Obie' -- {hpda, uwmcsd1}!sp7040!obie!wes "Happiness lies in being priviledged to work hard for long hours in doing whatever you think is worth doing." -- Robert A. Heinlein --
karl@triceratops.cis.ohio-state.edu (Karl Kleinpaste) (08/11/88)
wes@obie.UUCP writes:
While we're on the subject, is site tut.cis.ohio.edu passing ANY mail to
ANY other sites at all? Everything I've sent through them lately has
bounced, to the point where I re-ran pathalias with tut marked as
'dead.'
The machine tut.cis.ohio-state.edu has never performed UUCP mail
pass-through. It has only one UUCP neighbor, osu-cis, about six feet
away in the next cabinet with a cable between them.
osu-cis, in turn, has relatively few UUCP neighbors, the primary
non-local ones being pyramid and killer. We also speak to AT&T, as
one might guess, but since that's not useful for pass-through any
longer, we are marking it (DEAD), or, as Tim Thompson recommended to
us today, with the <att>(DEMAND) syntax, which defines the link as
terminal.
If the reason this concerns you is because you find that a lot of
!-path routed mail gets aimed through Tut and fails, I'm well aware of
it - I see a couple of them every day. But such things are invariably
due to replies generated to the Path: header of news articles, and
such paths simply don't work - they hit Tut requesting UUCP transfer
to some site with which we connect solely via NNTP, and they will at
best get rerouted via, say, Rutgers. If they work, you got lucky. If
they don't, they bounce. Don't generate replies to Path:.
I have no idea why Tut shows up in pathalias output anyway; our map
entry doesn't even list Tut except as a neighbor of osu-cis, cost
value LOCAL, and in the #U line.
--Karl