[Mailman-Users] Probe Message Problem

[Mark Sapiro]

Gary Kaplan wrote (accidently?) off list:

>I have no idea what  Marks response means. Can someone please help?
>
>-Gary
>
>Mark Sapiro wrote:
>
>> Gary Kaplan wrote:
>>
>>> One of my users informed me that she got this message in which a couple
>>> hundred address from the list were disclosed. I want to know if there 
>>> is a
>>> way to stop this. I don't want everyone who gets a probe message to see
>>> everyone's email address. I have noted where the address are listed 
>>> but left
>>> the actual addresses off.
>>>
>>>
>>>> This is a probe message.  You can ignore this message.
>>>>
>> <snip>
>>
>>>>
>>>> Received: from mailnull by washington.hostforweb.net with local 
>>>> (Exim 4.41)
>>>>     id 1C9XAp-000348-CZ
>>>>     for Upcnews-bounces at upc-online.org; Mon, 20 Sep 2004 19:02:19 -0400
>>>> X-Failed-Recipients:
>>>
>>>
>>> At this point it lists a couple hundred email addresses
>>>
>>> Then is says this:
>>>
>>>> Auto-Submitted: auto-generated
>>>> From: Mail Delivery System <Mailer-Daemon at washington.hostforweb.net>
>>>> To: Upcnews-bounces at upc-online.org
>>>> Subject: Mail delivery failed: returning message to sender
>>>> Message-Id: <E1C9XAp-000348-CZ at washington.hostforweb.net>
>>>> Date: Mon, 20 Sep 2004 19:02:19 -0400
>>>>
>>>> This message was created automatically by mail delivery software.
>>>>
>>>> A message that you sent could not be delivered to one or more of its
>>>> recipients. This is a permanent error. The following address(es) 
>>>> failed:
>>>
>>>
>>> and list the addresses again.
>>
>>
>> It appears that Mailman is appending the bounce message to the probe
>> and in this case there were "a couple hundred" subscribers in the
>> washington.hostforweb.net domain that bounced.
>>
>> In any case, this seems to be sufficient reason to not append the
>> original bounce mesage to the probe.
>>
>> I'm far from fluent in Python, but it looks like either
>>
>>  self.sendProbe(member, msg)
>> in the routine
>>  def registerBounce(self, member, msg, weight=1.0, day=None):
>> in Bouncer.py
>>
>> or
>>  outer.attach(MIMEMessage(msg))
>> in
>>  def sendProbe(self, member, msg):
>> in Deliverer.py
>>
>> are places where this can be addressed.
>>

OK, let me try again.

>From the Mailman-2.1.5/NEWS file:

    - The bounce processor has been redesigned so that now when an
      address's bounce score reaches the threshold, that address will
      be sent a probe message.  Only if the probe bounces will the
      address be disabled.  The score is reset to zero when the probe
      is sent.  Also, bounce events are now kept in an event file
      instead of in memory.  This should help contain the bloat of the
      BounceRunner.

What appears to have happened in your case is that you have "a couple
hundred" subscribers in the washington.hostforweb.net domain that all
bounced to the point of reaching the threshold causing a probe message
to be sent to each of the bouncing addresses. In at least one case and
maybe in all, the bouncing address was actually valid because the
probe reached the subscriber.

The problem here is that the bounce message that caused the score to
reach the threshold was appended to the probe, and it contained a list
of all the bouncing addresses in the domain.

I don't think it is necessary for Mailman to append the bounce message
to the probe message, and I think your experience is good reason not
to. Unfortunately, there appears to be no way in 2.1.5 to fix this
short of hacking the python code and I suggested a couple of places
where one could start looking if one wanted to do this.

As far as how often this might happen on your list is concerned, if all
or almost all the addresses are actually valid and do not normally
bounce, there was apparently some MTA transient situation that caused
these addresses to all be bouncing at once. If this is a situation
that might occur again, you might consider tuning your bounce
processing parameters to minimize the likelihood of valid addresses
reaching threshold. If, OTOH, almost all the addresses were in fact
bad, they are now gone and won't be "revealed" again.

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