[comp.mail.sendmail] Help with deferred mail bug

katinsky@schaefer.MATH.WISC.EDU (Matthew Katinsky) (03/31/89)

Eek!  A bug!  (No, its a design decision)

We have an AT&T 3B15 and several 3B2's all running system five unix.
What happens is the 3b15 (named schaefer) tries to send mail to a sys admin
on another 3B2 over ethernet.  Something goes wrong and it recollects
the message with most of its header, adds the following lines

   	----- Transcript of session follows -----
	554 qfAA09619: line 7: readqf(AA09619): bad line ""

   	----- Unsent message follows -----

then adds a new header and tries to mail it to the mail administrator, which
includes me among others.  It cannot mail it to us again, so it adds a new
header, the above lines again, and the recycled message and mails it again.
In the process, the mail message grows and grows.  When it has decided it
has had enough of this, it does manage to put it in my mailbox and I get this
huge mail message consisting mostly of headers.  But that's not all.  It then
sends it to me again and again, each time with an additional header and
"Transcript of session follows ..."

I try to read my mail and nothing happens (poor response time).  So I check the
size of my mailbox in /usr/mail and it is anywhere from 500K to 1 megabyte
full of the same recycled message.

I have been unable to detect what messages sets this off.  It is usually but
not exclusively mail from a daemon.  It is always mail routed through schaefer
(our connection to the outside world).  I am told that the error 

	bad line ""

refers to the fact that the message doesn't begin with a blank line as it
should.  But if this is a daemon sending it, shouldn't either it or the mail
program know to put the blank line in?   It seems ridiculous to me that the
mailer daemon would take an incorrectly formatted message, and send me another
incorrectly formatted message to tell me about it.

If anyone can tell me what might be wrong and/or what to do about it I would be
very grateful.  Please e-mail to katinsky@math.wisc.edu  and I'll post
responses.

---An abridged example found in my mailbox (originally over 5000 lines)---

From Mailer-daemon Wed Mar 29 12:34:38 1989
Received: from schaefer.math.wisc.edu (schaefer.math.wisc.edu.ARPA) by slichter.math.wisc.edu;
          id AA14422; 4.12/14; Wed, 29 Mar 89 12:34:38 cst
Received: by schaefer.math.wisc.edu;
          id AA09803; 4.12/14; Wed, 29 Mar 89 12:16:46 cst
Date: Mon, 27 Mar 89 00:50:18 est
From: Mail Delivery Subsystem <Mailer-daemon>
Subject: Returned mail: Deferred: No such file or directory
Message-Id: <8903291816.AA09803@schaefer.math.wisc.edu>
To: Mailer-daemon

   ----- Transcript of session follows -----
554 qfAA09619: line 7: readqf(AA09619): bad line ""

   ----- Unsent message follows -----
Return-Path: <Mailer-daemon>
Received: by schaefer.math.wisc.edu;
          id AA09619; 4.12/14; Wed, 29 Mar 89 11:51:48 cst
Date: Mon, 27 Mar 89 00:50:18 est
From: Mail Delivery Subsystem <Mailer-daemon>
Subject: Returned mail: Deferred: No such file or directory
Message-Id: <8903291751.AA09619@schaefer.math.wisc.edu>
To: Mailer-daemon

   ----- Transcript of session follows -----
554 qfAA09448: line 7: readqf(AA09448): bad line ""

	.
	.
	.
	[More of the same]
	.
	.
	.

   ----- Unsent message follows -----
Return-Path: <Mailer-daemon>
Received: by schaefer.math.wisc.edu;
          id AA03294; 4.12/14; Mon, 27 Mar 89 01:20:46 cst
Date: Mon, 27 Mar 89 00:50:18 est
From: Mail Delivery Subsystem <Mailer-daemon>
Subject: Returned mail: Deferred: Connection refused
Message-Id: <8903270720.AA03294@schaefer.math.wisc.edu>
To: Mailer-daemon

   ----- Transcript of session follows -----
554 qfAA02581: line 7: readqf(AA02581): bad line ""

   ----- Unsent message follows -----
Return-Path: <uucp>
Received: by schaefer.math.wisc.edu;
          id AA02581; 4.12/14; Mon, 27 Mar 89 00:50:18 est
Date: Mon, 27 Mar 89 00:50:18 est
From: The uucp Administrator <uucp>
Message-Id: <8903270550.AA02581@schaefer.math.wisc.edu>
To: uucp
Subject: cores

/usr/spool/uucp/outbac/core
/usr/spool/uucp/mpss/core
/usr/spool/uucp/dfmp1/core
/usr/spool/uucp/uwvax/core
/usr/spool/uucp/core

-----------------------------------------------------

katinsky@math.wisc.edu