[Mailman-Users] bounce processing settings

[Mark Sapiro]

Christopher Adams wrote:

>There seems to be lots more spam/bounce activity lately and I 
>am wondering if I should consider changing the way bounce behaves.


If you're thinking about 'unrecognized bounces' due to spam being sent
to the list-bounces address, your bounce processing settings won't
affect this.

Basically, spam gets involved in bounce processing in only a couple of
ways. When spam is sent directly to the list-bounces address, this
essentially always is an unrecognized bounce, because it doesn't look
like any known delivery status notification (DSN).

When spam is sent to the list address, and the 'post' is rejected or
held (say because the spammer is not a list member and the list
rejects or holds non-member posts), the reject/held notice often
bounces because the address that the notice is returned to is not
deliverable. This results in a legitimate bounce which may be
unrecognized because the DSN is not one that Mailman can parse, or it
is recognized as a bounce for a non-member address and ignored.

The list's bounce settings (other than
bounce_unrecognized_goes_to_list_owner) have no effect on any of this.

>I 
>could easily change all lists globally using withlist, but I thought 
>that I would first see what others are doing.
>
>For 'bounce detection sensitivity', I  am currently using:
>
>5, 7, 3, 7


These settings are the defaults because they are reasonable settings
for most lists. If you are sensitive about sending too many messages
to dead addresses (AOL, e.g. dings you for this), and you aren't too
concerned about disabling people for 'full mailboxes', etc. you could
lower the threshold.

Also, for lists with low activity - e.g. one or two posts per week, you
probably want to both lower the threshold, and extend the 'stale
after' time.

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