[Mailman-Users] help with rather time-sensitive dilemma??

[Amanda]

Well, I'll tell you what I know so far....

J>  What happens if you run mail or mailx from the command line and
J>  send a message to the owner address (testlist-owner at tux)?  Does it
J>  get delivered?

Using "mail" from the command line to send mail to testlist-owner does not
arrive. Mailx? was ist das?

J>  What happens if you telnet to port 25 on the localhost and
J>  hand-write a message to the owner address?  Does it
J>  get delivered?

Same thing. qmail's accepting it, but it never arrives.

J>  Repeat the last with the FQDN for the host.

No change.

J>  What happens if you use 'sendmail -t' to send a message to the
J>  owner?  Does it get delivered?

Either tux can't do that, or I'm doing it wrong. (Money's on the latter.)
"sendmail -t" sends it into a tailspin. I have to CTRL+C to get out of it.

J>Which delivery model are you using under Mailman: sendmail or
J>SmptDirect?  (recommend SmtpDirect)

Yep. SMTPDirect.


B>Word of warning: I've never used qmail.

Funny! Neither have I, up until about seven weeks ago. ::crosseyed look::

B>Try to use Python's smtplib module directly to mimic the same calls
B>  Mailman is making to see what kind of failure or return value you're
B>  getting.

Erfgh! I know I look real smart with my cute little red-haired ponytail
and all, and have learned enough perl to get me out of a few paper bags,
but that's about as much of a programmer-type-person as I is. Perhaps you
could take the time to use small words and tell me how to do that? (I know
that's about as much fun as untying knots in Christmas lights...)

B>Try adding a syslog() call to SMTPDirect.py's process() function,
B>  inside the "if code >= 500 and code <> 552" clause to log exactly
B>  which error codes you're receiving.  Something like:
B>
B>  syslog('debug', 'Received error code %d for recipient %s' % (code,
recip))
B>
B>  Then tail the logs/debug file to see which error codes your
B>  getting.  Then try to figure out why qmail is returning these error
B>  codes. ;)

soooo.... if I poke around in smtpdirect.py I will find a section that
looks like process(put yer left foot in, put yer left food out, if code
>=500 and etc, do the hokey-pokey) and at some point in that section I
should put in the line you mention above? Will it be really painfully
obvious where it should go?

J> I've got a quiet bet running that QMail is configured to
J> either refuse/error connections from 127.0.0.1 or from the
J> local IP.

B>I think that's a darn good bet.  Note that if Amanda hasn't otherwise
B>customized them, the defaults for SMTPHOST or SMTPPORT are 'localhost'
B>and 0 respectively with port=0 meaning "use the standard smtp port",
B>i.e. she's attempting to connect to 127.0.0.1 port 25.

qmail depends entirely on xinetd to handle smtp connects. xinetd.conf does
not say anything, either in defaults or in service smtp, about only_from
or no_access. I'm under the impression that that means anyone can get to
it - including local loopback addresses.

I've got some other things in the logs that I think might be of interest
(or not...):

/var/log/servicelog (which is where xinetd does its thing) reports, in
relevance to my examples yesterday:
01/7/10 at 15:17:01: START: smtp pid=6302 from=127.0.0.1
01/7/10 at 15:37:03: START: smtp pid=6421 from=127.0.0.1
01/7/10 at 15:37:03: START: smtp pid=6422 from=127.0.0.1
01/7/10 at 15:37:04: START: smtp pid=6423 from=127.0.0.1

And the telnet sessions I ran to the smtp server look like this:
01/7/11 at 12:50:05: START: smtp pid=10165 from=127.0.0.1
01/7/11 at 12:52:30: START: smtp pid=10176 from=(tux's real ip)

And the telnet sessions look like this in /var/log/syslog (which is where
qmail does its thing):
Jul 11 12:49:26 tux qmail: 994880966.397577 new msg 114919
Jul 11 12:49:26 tux qmail: 994880966.397875 info msg 114919: bytes 305
from <amanda>
qp 10092 uid 507
Jul 11 12:49:26 tux qmail: 994880966.414473 starting delivery 9: msg
114919 to local
testlist-admin at tux.[foo.bar]
Jul 11 12:49:26 tux qmail: 994880966.414749 status: local 1/10 remote 0/20

Jul 11 12:49:27 tux qmail: 994880967.566871 delivery 9: success:
did_0+0+1/
Jul 11 12:49:27 tux qmail: 994880967.567248 status: local 0/10 remote 0/20

Jul 11 12:49:27 tux qmail: 994880967.567352 end msg 114919
Jul 11 12:50:00 tux CROND[10162]: (root) CMD (   /sbin/rmmod -as)
Jul 11 12:50:00 tux CROND[10163]: (mailman) CMD (/usr/bin/python -S
/home/mailman/cron/qrunner)
Jul 11 12:50:00 tux CROND[10164]: (mailman) CMD (/usr/bin/python -S
/home/mailman/cron/gate_news)
(...)
Jul 11 12:55:17 tux qmail: 994881317.698711 new msg 114919
Jul 11 12:55:17 tux qmail: 994881317.699044 info msg 114919: bytes 313
from <amanda>
qp 10177 uid 507
Jul 11 12:55:17 tux qmail: 994881317.726899 starting delivery 10: msg
114919 to
local testlist-admin at tux.[foo.bar]
Jul 11 12:55:17 tux qmail: 994881317.727140 status: local 1/10 remote 0/20

Jul 11 12:55:18 tux qmail: 994881318.924386 delivery 10: success:
did_0+0+1/
Jul 11 12:55:18 tux qmail: 994881318.924767 status: local 0/10 remote 0/20

Jul 11 12:55:18 tux qmail: 994881318.933131 end msg 114919

I really appreciate the help... and if you have further suggestions I am
all ears!

=)
Amanda