[Mailman-Users] Mailman speed

[Ian A B Eiloart]

--On Thursday, July 8, 2004 3:26 pm +0200 Brad Knowles 
<brad.knowles at skynet.be> wrote:

> At 11:18 AM +0100 2004-07-08, Ian A B Eiloart wrote:
>
>>  You can let the MTA do the VERP for you. That is faster than letting
>>  mailman do it. <http://www.exim.org/howto/mailman21.html#verpex>
>
> 	That may be true for Exim, but perhaps not for other MTAs. Assuming
> that's true, it's still going to slow down the MTA quite a bit.

Well, the overall system performance will be intermediate between no VERP 
and letting Mailman do the VERP.

Mailman would only need to pass one copy of the message to the MTA (subject 
to max_recipient rules on both).

The MTA doesn't need to VERP local deliveries. If many of your recipients 
are local, as they would be on a corporate or university system for 
internal lists, then there is a large saving here.

The MTA does need to VERP remote deliveries, and that means making callouts 
to remote machines for each recipient. However, exim will use a single 
connection for each remote host - rather than one per recipient. So, exim 
has to send the data several times, but it doesn't have to read the data 
several times.

> 	Moreover, I question the value of this within the context of Mailman --
> since Mailman didn't do the VERP on outbound, I wonder whether or not it
> will be able to make appropriate use of the extended information on the
> incoming bounce.

Yes, it will. There would not have been much point doing it otherwise. The 
ONLY change to the Mailman configuration there is VERP_DELIVERY_INTERVAL. 
The exim recipe does the

-- 
Ian Eiloart
Servers Team
Sussex University ITS