Frank Bures wrote:
> [log in to unmask] wrote:
>> Frank Bures wrote on 04/02/2009 09:31:49 AM:
>>
>>> Lately I have been banging my head against the following problem.  I
>> would
>>> very much appreciate any ideas, no matter how wild they might seem.
>>>
>>> What happens:  I have mostly local subscribers, but some are from
>> outside
>>> domains like gmail, hotmail and yahoo.  One of the outside domains is
>> our
>>> university central mail system (our department is running its own mail
>>> system).  Subscribers, whose addresses are in the university central
>> mail
>>> system intermittently receive list posts duplicated.  I traced the
>> problem
>>> in the logs and here is an example of what happens:
>> It would help if the user would send you the full headers to both messages
>> that were duplicates.  Look at where they are the same, and where the
>> difference occurs, mostly in the Received headers.
>>
>
> I already know where they differ.  LISTSERV simply intermittently decides
> to resend one whole block of addressees some 10 minutes later.  The
> re-sending always occurs after approx. the same time interval.  Yet, there
> is no mention of that in listserv log.
>
> Anyone has seen that?
>
> Cheers
> Frank
>

Further to my query - detailed investigation revealed a disturbing behaviour:

When LISTSERV passes the message to sendmail, it creates blocks of
addressees containing up to 100 entries.  In my particular case, I had 8
such blocks for some 750 subscribers.

When the duplicate mailing occurs, LISTSERV passes one (and just one) of
those blocks to sendmail again, some 10 minutes after the original mailing.
The block always contains external addresses.  They do not have to be
related to our central university mailing system.  Duplicate mailings
occurs intermittently to yahoo and other universities as well.  In other
words, any addresses that are in that single block of 100 entries will be
re-mailed.

As I wrote before, there is no trace of that second LISTSERV single block
mailing in listserv log, only in maillog.

Cheers
Frank

--

Frank Bures, Dept. of Chemistry, University of Toronto, M5S 3H6
[log in to unmask]
http://www.chem.utoronto.ca
PGP Public Key:
http://subkeys.pgp.net:11371/pks/lookup?op=index&search=Frank+Bures

Please, reply to the above address only.  <lisfrank> does not accept replies.