[Mailman-Users] can't get mailman and exim4 talking on my debian 3.1

[Mark Sapiro]

Please include the mailman-user's list in replies.

No individual on the list knows everything.

Individuals make mistakes.

Individuals have other lives that take them away from mailman and/or
email for days or weeks at a time.

The list is a community which can often offer better and/or more timely
response than an individual. At worst, the individual will see your
list post and reply to it anyway.

If nothing else, a coherent thread in the archive can help others long
after the fact.



Dr. Scott S. Jones wrote:
>
>> >> Did you add the line
>> >> 
>> >>         self.__conn.set_debuglevel(1)
>> >> 
>> >> to Mailman/Handlers/SMTPDirect.py and restart Mailman and then look in
>> >> Mailman's 'error' log?
>
>I did add that line, but I will check again to make sure it is there
>correctly. 
> 
>> That's why you see the "No such list" errors. Someone has tried to
>> access those lists - in the case of 'test75' via the admin interface,
>> in the case of the others, via perhaps listinfo, private or some other
>> script that doesn't log it's name.
>> 
>> Anyway, the only messages relevant to your issue are "Cannot connect to
>> SMTP server localhost on port smtp". I don't know why there aren't
>> also debug prints from the Python smtplib if the
>> 
>>         self.__conn.set_debuglevel(1)
>> 
>> line was added to SMTPDirect.py per the FAQ. Basically however, this
>> says that in these instances at least, SMTPDirect was unable to
>> connect at all to 'localhost' port 25.
>
>I can only assume that the 'someone' to whom you refer, must be an outside
>entity attempting to email to those lists which are now gone.


Emailing a non-existent list can produse the error log entry

  'post script, list not found: %(listname)s'

But only if there are aliases or something left in the MTA that still
delivers to the list. In your case, Exim won't attempt to deliver to a
non-existent list, and in any case, your error log entries don't look
like the above.

The entries in your log are caused by someone or something (anyone,
anything, including a web crawler) accessing a URL like
http://example.com/mailman/<script>/<list> where <list> doesn't exist.

In one of the three cases, the script was 'admin', because 'admin'
identifies itself in that message. In the other two cases, the message
was the generic "No such list <list>: message which is issued py the
scripts admindb, confirm, edithtml, listinfo, options, private, rmlist
and subscribe.

If you want to know more about what script was invoked and from what
IP, consult your web server log for the times in question.


>If I knew the
>domain of your server, where your lists are hosted, and I attempted to mail
>to "nonexistant_list at mark_sapiro.org" ... is it safe also to assume that
>such email to that nonexistant list would generate a similar type error
>message in your log file?


No, it would be rejected as undeliverable by the incoming MTA (or
perhaps just discarded depending on how my incoming mail is delivered
to Mailman).


>If so, then I understand what you are saying about
>my lists which I removed. 
>
>Now that I know that SMTPDirect won't connect to 'localhost' port 25, how do
>I proceed? I guess I am having difficulty understanding how simply upgrading
>mailman or exim4....when I opted to keep my existing configuration files,
>would so badly hose my Mailman installation as to make it unworkable. 


That does seem unusual, but that's what we're trying to figure out.


>Is it safe to cut back on one's log files, for example with
>/var/log/mailman/error, keeping only the past two weeks of entries? 


I use logrotate to rotate the logs weekly and keep for prior
generations in addition to the current. If you do this, don't forget
to tell logrotate to do 'bin/mailmanctl reopen' after rotating.


>I am including a few other files to help you help me figure this out. 



I have snipped this in case there is anything in it that you don't want
to go to the list.

I think you said that you can successfully 'telnet localhost 25' and
get an appropriate '220' from Exim, and that you successfully ran the
various tests in FAQ 6.14, but that message was off list so I'm not
sure exactly what you said. Did you see the paragraph

<quote>
If the various diagnostic snippets above don't fail, first make sure
you are running them as the 'mailman' user and not as root as this can
make a difference. Also see
<http://mail.python.org/pipermail/mailman-users/2005-May/044972.html>
and
<http://mail.python.org/pipermail/mailman-users/2005-May/044742.html>
for more complete tests, and if these work, make sure you don't have
different (wrong) settings for SMTPHOST and/or SMTPPORT in mm_cfg.py.
</qoute>

in the FAQ and do the tests in the referenced posts as the mailman user?

In any case, your problem doesn't seem to be the one in FAQ 6.14 as
that is a hard failure, and your's seems more intermittent,
particularly considering your "slow processing" thread, however the
tests in the referenced posts may still be useful.


>Scott
>
>p.s. I am very grateful for your help here. I am sorry I am at times brain
>dead about how to proceed. 


It's OK - we all have blind spots from time to time.

-- 
Mark Sapiro <msapiro at value.net>       The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan