[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Emacs sends mail, but doesn't show-up in exim.



On Mon, 14 May 2001 at 11:38:48 -0400, Rob Mahurin wrote:
 [snip]
 >Try sending a message to a both to a disappearing and
 >non-disappearing address.  The disappearing address, after
 >delivery, should show up with a "D" in the mailq output,
 >while the non-disappearing address prevents the mail from
 >being erased from the queue.  Then you should be able to
 >see what happened from the logfile.
 [snip]

I have also been doing a lot of reading:
 - /usr/share/doc/exim/oview.txt.gz
 - /usr/share/doc/exim/spec.txt.gz

To answer one of my own questions previously posted, the way
to see log messages for email messages that have already been
delivered is to use ``exigrep'' ...for example:

  # exigrep aol /var/log/exim/mainlog

And ``exicyclog'' causes log files to be refreshed.
Thank you for explaining the syntax for
mailq/head/cut/tail commands.  OK, i tried sending a message
to both a disappearing and a non-disappearing address. I knew
that "@yahoo.com" would get to its destination, but "@aol.com"
would not:

  $ echo "TEST." | mail rcymbala@yahoo.com Hot49erfan@aol.com

Here it is in the queue:

  # mailq
   0m   307 150UOr-0001mF-00 <cymbaLa@Lafn.org>
        D Hot49erfan@aol.com
          rcymbala@yahoo.com

Here it is in the main log (``exicyclog'' was used to flush old
messages away):

   # exigrep aol /var/log/exim/mainlog
   +++ 150UOr-0001mF-00 not completed +++
   2001-05-17 13:29:33 150UOr-0001mF-00 <= cymbaLa@Lafn.org U=cymbala P=local
S=289
   2001-05-17 13:29:33 150UOr-0001mF-00 == rcymbala@yahoo.com routing defer
(-42): retry time not reached
   2001-05-17 13:29:33 150UOr-0001mF-00 == Hot49erfan@aol.com R=smarthost
defer (-1): 
   2001-05-17 13:29:33 150UOr-0001mF-00 ** Hot49erfan@aol.com: retry timeout
exceeded
   2001-05-17 13:29:34 150UOr-0001mF-00 Error message sent to cymbaLa@Lafn.org

After doing enough reading, it occurred to me that the reason
why email to "Hot49erfan@aol.com" (and select others) was marked as
"D" for delivered (per man-page extract you posted in most recent
reply) but was "disappearing" had to do with Exim's databases.
It has these databases:
 - retry
 - reject
 - wait-*   (e.g., wait-smtp)
 - serial-*

Those databases remember what has happened in the past.  The command
to view those databases is "exim_dumpdb".  So, i dumped:

  # exim_dumpdb /var/spool/exim retry

Lo and behold, the "problem" domains (like "@aol.com") appeared with
the strange message "unknown error" (or something like that).

I don't have the exact error message from the retry database because of:

  # exim_tidydb -t 0d /var/spool/exim retry

That command deleted all historical information (thus, I don't have
wording of the error message).

That did it!!!!!  I mean, fixed the problem.

After running tidy-db (with time set to 0d, or zero days)
the retry database no longer gets in the way.  In other words:

  # mailq
   0m   307 150UOr-0001mF-00 <cymbaLa@Lafn.org>
          Hot49erfan@aol.com
          rcymbala@yahoo.com

Now email to Hot49erfan@aol.com has _NOT_ yet been delivered, but will
be delivered once I establish a PPP connection using ``pon''.

Two observations:

1.
  file://localhost/usr/share/doc/exim/oview.txt.gz
  "Exim is not designed for storing mail for dial-in hosts."

2.
  file://localhost/usr/share/doc/exim/spec.txt.gz
  "41. INTERMITTENTLY CONNECTED HOSTS
       Exim was designed for use on permanently connected hosts."
and,
  "41.2 Exim on the intermittently connected host"

3.
  ``man exim_tidydb''
   "It is suggested that it be run periodically on all three
    databases, but at a quiet time of day, ..."

Now there's just one question remaining in my mind... is Debian
configured to run exim_tidydb periodically?  In my situation
(the dial-up laptop) the whole retry/reject database thing is a hassle
that doesn't really increase efficiency.  I'm probably going to put
exim_tidydb (with time 0d) into my daily cron jobs just to avoid
the original problem: disappearing emails.

Final note: exigrep reported that error messages were being
sent upon "retry timeout exceeded" but the destination domain
("@lafn.org") was one of those domains in the "retry timeout
exceeded" group.  Thus, error notifications were also being
rejected, giving the "disappearing message" effect.

Here's the original post:
  http://lists.debian.org/debian-user-0105/msg00755.html


THANKS Michael, Alan and Rob!

Robert
--
     Robert Cymbala cymbaLa@Lafn.org ...or... rCymbala@yahoo.com
    QUEBEC CITY -- The leaders of 34 Western Hemisphere nations...
  promised to create a hemisphere in which `no one is left behind.'
	      -- April 23. JAMES GERSTENZANG, L.A. Times



__________________________________________________
Do You Yahoo!?
Yahoo! Auctions - buy the things you want at great prices
http://auctions.yahoo.com/



Reply to: