> -----Original Message-----
> From: Manuel Lemos [mailto:[EMAIL PROTECTED]]
> On 01/26/2003 08:22 PM, Mark Charette wrote:
> >>-----Original Message-----
> >>From: Manuel Lemos [mailto:[EMAIL PROTECTED]]
> >>That number of messages is small and with today's servers you can queue
> >>that volume of personalized is less than a minute. Your problems will
> >>start as you enter in the tens of thousands of recipients, not because
> >>qmail won't handle it but because that many message in the queue will
> >>prevent the incoming messages to arrive in time.
> > ???
> >
> > There are two separate processes here - the outgoing connections to the
> > remote SMTP servers and the incoming connections to your own
> SMTP servers.
> > Tuning qmail involves setting the number of outgoing queues so as not to
> > overload your communications line.
> >
> > I _personally_ have sent more than 300,000 messages with no "resting"
> > mechanism - doing that means you haven't tuned your qmail system for the
> > load.
>
> The problem does not have to do with SMTP but with qmail local queue
> handling. If you send messages to many users and the bounce address is
> set to some address handled by the same server, you will be starting to
> get bounces to the same machine and thus the same qmail server. The way
> to solve this is to put the bounce address point to some address that is
> handled by a different server or even to nowhere. Using a spare server
> for the actual delivery reduces that problem too.

It has everything to do with qmail local queue handling.

I have had no problems when both the VERP address and sending machine are
the same, providing I tune the qmail queues to a reasonable number. VERP is
very useful to handle bounces, no separate machines are required, you just
have to think about your bandwidth.

The way to solve this is use your intelligence. A stop/start sceme as you
originally proposed is a kludge at best to an easily managable problem.

Mark C.





-- 
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php

Reply via email to