[Mailman-Users] handling mail undeliverable to list owners

[Richard Haas]

Greetings.

We have a moderately sized Mailman installation (+2000 mailing lists) which
recently upgraded a long-standing Mailman v2.1.5 service to v2.1.14. The
upgrade went quite well, but since the upgrade the list owners of the
DEFAULT_SITE_LIST (i.e. "mailman") mailing list are seeing a marked increase
in the number of undeliverable emails to (other) list owner addresses, which
the "mailman" list owners did not see with the previous version/installation
of the software. While unlikely, they may have been a customization in
either Mailman or Postfix that was lost in the transition; the logic in
~/Mailman/Message.py which sends list owner bounces to DEFAULT_SITE_LIST
seems unchanged since 2.1.5 (at least).

Let me be clear: these are valid bounces from some list owner addresses that
do not deliver, because a person who was a list owner has departed the
organization and the mailbox to which a list owner address routes no longer
exists. These messages not errors in Mailman.

Often, a single list owner address will bounce to the DEFAULT_SITE_LIST,
while other valid list owners remain available for that same list -- in that
case, DEFAULT_SITE_LIST members need take no action. When there's just one
owner for a list and that address bounces, the list is orphaned (which is
good to know), but we have other means to reclaim unused lists (periodic
reregistration), and the administrative notice which bounced is frequently
not something that would be posted to the list anyway (e.g. spam, erroneous
out-of-office mail, spoofed sender, etc.) -- particularly if the list has
fallen into disuse. It would be nice to only see the bounces when there are
no other list owners (of the same list) to receive them.

Are there any Mailman configuration options which adjust the handling of
bouncing list owner emails, short of customizing the Messages.py or
adjusting filters on the DEFAULT_SITE_LIST policies? None leap out in
Defaults.py.

Thanks for any insight.

--
 Richard Haas <rhaas at rhaas.us>
--