[Mailman-Users] Throttling output

Dave Stern dave at umiacs.umd.edu
Tue Jun 13 19:22:27 CEST 2006

On Tue, 13 Jun 2006, Steve Campbell wrote:

> Bernd,
> ----- Original Message -----
> From: "Bernd Petrovitsch" <bernd at firmix.at>
> To: <mailman-users at python.org>
> Sent: Tuesday, June 13, 2006 9:17 AM
> Subject: Re: [Mailman-Users] Throttling output
>> On Tue, 2006-06-13 at 09:06 -0400, Steve Campbell wrote:
>> [...]
>>> Does anyone have a suggestion for throttling?
>> Just store the 22.000 outgoing mails in the mail queue (every decent MTA
>> should be able to do that unconditionally) and wait for the next queue
>> run?
> This originally sounded like a "too-simple-to-work" idea, but after thinking
> about it, it may be the best. If you mean QueueGroups with a separate runner
> for this particular list, then the milter would handle the throttle of the
> runner and relieve Mailman of the outgoing problems. It would really give me
> a little more control because of the sleep time of the queue runner along
> with the milter throttle. This will be a multiple-list server and only one
> or two lists would require the slowdown/pacing.
> Sounds great. In fact, this is sort of what the Milter people suggested
> without the milter, although I didn't see where the throttling would occur.
> Thanks
> Steve

Actually, that's exactly what we did. We had a tiny sun ultra10 and
ended up supporting a list of some 12,000 people. Aside from the 
standard tuning you'll read about in the Sendmail Performance Tuning
book by Christenson, I set up multiple queues limiting the maxquerunsize.

'Course now we're turning to linux and are considering using postfix 
instead but the theory is the same.

  =-=-=-=-=-=-=-=-=-=-  generated by /dev/dave -=-=-=-=-=-=-=-=-=-=-=-=
  David Stern                                    University of Maryland
            Institute for Advanced Computer Studies

More information about the Mailman-Users mailing list