[comp.mail.sendmail] I refuse to talk to myself?!?!

fuhrman@b.coe.wvu.wvnet.edu (Cris Fuhrman) (01/06/89)

I recently posted a question to comp.sys.sun about setting up mail on a 
a cluster of 10 suns, and the replies I got were just the trick.  Local
mail works perfectly now.  Thanks for the help with your replies!

The problem that now exists has to deal with incoming mail.  It seems that
if I send a message to the mailhost sun (called suna or 
suna.coe.wvu.wvnet.edu) two things happen:

 1)  If mail is addressed to, say fuhrman@suna (from another machine
     with nameserver resolver capabilities), all works fine.

 2)  If mail is addressed to, say fuhrman@suna.coe.wvu.wvnet.edu (fully
     specifying the hostname), the mail arrives at suna, but then 
     suna thinks that the address is foreign, and forwards the mail
     on to the MAJOR RELAY HOST in sendmail.cf (which happens to
     be itself) - et voila!  The following error message in mail from 
     MAILER-DAEMON:

     >>> HELO suna.coe.wvu.wvnet.edu
     <<< 553 suna.coe.wvu.wvnet.edu I refuse to talk to myself
     554 fuhrman@suna.coe.wvu.wvnet.edu... Service unavailable 

    I have tried changing the MAJOR RELAY HOST to be our VMS machine
    running CMU TCP/IP with nameserver.  All that happens here is that
    suna forwards the mail to the vax, and the vax sends it right back.
    The mail bounces about 15 times until both systems detect a loop
    and then it's canned - all because it doesn't recognize suna as
    being suna.coe.wvu.wvnet.edu.
 
Ok.  So I just want to know how sendmail resolves the equivalence of the
two names suna and suna.coe.wvu.wvnet.edu.  BTW, these two names _are_ on
the same line in /etc/hosts.  Also, the mailhost is also the loghost for
the yellow pages.  I have played and played with sendmail.cf.  What kind
of Black Magic is this and how can I get some info?  I have had compiler
classes before, and I think I'm capable of comprehending the syntax and/or
rewriting rules for the internet/uucp addresses.  I just don't know where
to look for info.

Please reply to the list or directly to manager@a.coe.wvu.wvnet.edu.  

All assistance would be *greatly* appreciated,

-Cris
-------------------------------------------------------------------------
| Cris Fuhrman             | manager@a.coe.wvu.wvnet.edu (129.71.12.1)  |
| Comp Sys Manager         | un020312@wvnvms.bitnet                     |
| G-40B Eng Science Bldg.  | "Just say NO to commercial file x-fers."   |
| WVU Morgantown, WV       |              - Kermit the Frog             |

avolio@decuac.dec.com (Frederick M. Avolio) (01/06/89)

You must let sendmail know in the configuration file that host.domain is
the same as host which is the same as YOU.  No havig your sendmail.cf file
in front of me I can't tell you where to fix it but it should be
somewhere near where you handle anything.yourdomain.  Stick a line in
to resolve that just hostname (if you want) or put a rule in S0 before the
rule that sends host.domain to the tcp mailer to look for localhost.domain
and send it to the local mailer.

Fred

deke@socrates.ee.rochester.edu (Dikran Kassabian) (01/07/89)

The sendmail 'built in check' causing this is protecting you from an
infinite loop.  What I think you need to do is check the rules in your
configuration file that recognize 'myself' in the address.  You have to
rewrite that address to be just the user portion so that things can
resolve to the local mailer and can be delivered.

trivial example assuming the macros $w and $D have their usual meaning as
fully-qualified hostname and mail-domain respectively:

R$-@$D         $1                 rewrite stuff@mydomain to stuff only
R$-@$w         $1                 rewrite stuff@mymachine to stuff only

These are off the top of my head but you get the idea. Maybe the rules are
incorrect or missing.  Maybe the macros are incorrect or missing.  Maybe
the ruleset is never invoked or improperly formed.  I'd check that class
of things first.

I hope that this is helpful.  Good luck!


      ^Deke Kassabian,   deke@ee.rochester.edu   or   ur-valhalla!deke
   Univ of Rochester, Dept of EE, Rochester, NY 14627     (+1 716-275-3106)