Got lots of suggestions, for answers, thanks. Found the problem myself not
long after my posting appeared in my mail box. Here's the problem:
> Configuration: mail relay host/server running SunOS 4.1.2
> and another running Solaris 2.3 (problem happens
> no matter which relay host is used)
> mail client running Solaris 2.3
> Problem: We have a particular user such that mail to him:
> is delivered just fine.
> Mail to our original user with the bouncing problem
> with any other machine in our mail domain following
> the @ sign is delivered with no problem.
> Here is the message from Mailer-Daemon:
> > From Mailer-Daemon@cps.msu.edu Tue Jan 18 15:37 EST 1994
> > Date: cc
> > From: email@example.com
> > Subject: Returned mail: unknown mailer error 1
> > Subject: test
> > To: <firstname.lastname@example.org>
> > ----- Transcript of session follows -----
> > assign(debug)=
> > Cannot give -f and people to send to.
> > Usage: mail -eiIUdFntBNHvV~ -T FILE -u USER -h hops -r address
> > -s SUBJECT -f FILE users
> > 554 <ni@ss36>... unknown mailer error 1
The problem description turned out to be inaccurate...mail to any user@mailclient
bounced, if the mail should have been delivered using Mlocal. This was hidden by
the fact that the other test accounts I used forwarded their mail, whereas the
bouncer did not.
The problem was further complicated by the fact that I had been testing the
configuration file incorrectly. The mailclient mounts /var/mail, so
I had set the OR option, but when testing the file, ran addresses through
rule set 0, not rule set 33. I had searched the AnswerBook for references to
this new rule set when I saw it in the 2.3 configuration file, but as I could
find no references, decided that it couldn't be all THAT important :-)
Anyway, rule set 33 is used instead of (?) rule set 0 in the cf file when the
option R is set. So the mail client was trying to deliver the mail using
Mlocal, but /bin/mail was failing because (I assume) the directory is nfs mounted.
I GLAD it failed, could have been pretty darn ugly if it hadn't.
This was not a fun exercise. Did I miss something in the AnswerBook and
this rule set IS documented? If not, I think Sun needs a wake up call. Mail is
a bit too important for changes like this to be poorly documented. My next
big project is Sendmail V8 domain wide...life will hopefully be much nicer
for me after that.
Thanks to everyone who replied. I appreciate the time and thought you gave
my problem. Things like this are so hard to debug long distance.
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:08:55 CDT