...

...

...

...

"J" == J C Lawrence <claw@kanga.nu> on Tue, 11 Nov 2003 09:05:45 -0500 writes:

J> On Tue, 11 Nov 2003 10:52:07 +0100 
J> Martin Maechler <maechler@stat.math.ethz.ch> wrote:

>> My problem is that I get about 50 "Uncaught bounce notification"s per
>> day, only few of which are spam/virus related. 

J> Turn on VERP for a few days.

J> -- 
J> J C Lawrence                
J> ---------(*)                Satan, oscillate my metallic sonatas. 
J> claw@kanga.nu               He lived as a devil, eh?		  
J> http://www.kanga.nu/~claw/  Evil is a name of a foeman, as I live.

Thank you, "J", for the suggestion, but that seems not an option for two reasons

1. we (currenly must) use sendmail and that doesn't seem to support VERPing -- or at least we haven't found docu on that

2. VERP will probably cost quite a bit of CPU/memory/disk-IO resources on the mail servert. Currently that server almost constantly runs on load 2.5 -- 3 (because of anti-virus / anti-spam / mailman ..)

OTOH, most of these "uncaught bounce notification" can quite easily resolved by a human, i.e., the subscriber's e-mail can very often be found in the bounce. E.g., if they resend the full headers, I can find the proper subscriber's address in a "Received:" header such as

Received: from hypatia.math.ethz.ch (root@hypatia.ethz.ch [129.132.58.23]) by cryforhelp.mr.itd.umich.edu (umich) with ESMTP id hAB7NrHc014893 for <pgreen@umich.edu>; Tue, 11 Nov 2003 02:23:53 -0500

(hypatia is our own mail server)

I.e., I think mailman's analysis of the bounce message leaves room for improvement. Can you tell me where in the python code this happens (so we might consider improving it here)?

Best regards, and thank you all for mailman!

Martin Maechler <maechler@stat.math.ethz.ch> http://stat.ethz.ch/~maechler/ Seminar fuer Statistik, ETH-Zentrum LEO C16 Leonhardstr. 27 ETH (Federal Inst. Technology) 8092 Zurich SWITZERLAND phone: x-41-1-632-3408 fax: ...-1228 <><