[Mailman-Users] Throttling output

[campbell at cnpapers.com]

Quoting Brad Knowles <brad at stop.mail-abuse.org>:

> At 4:43 PM -0400 2006-06-14, Steve Campbell wrote:
> 
> >>  This is an MTA configuration issue.  There's nothing here to "fix"
> >>  within Mailman.
> >
> >  I think I just discovered that Mailman does deliver to the mqueue
> directory.
> 
> 	It has it's own internal queue directory, yes.  But it does not 
> deliver directly to the MTA mqueue directory, no.  Instead, it hands 
> the message off to the MTA that it is configured to use -- usually 
> port 25 on the "localhost".

It eventually ends up there in mqueue, I saw, but I couldn't see how. I did know
that there is a nice array defined in Defaults.py, and the input and output
queues are in and out respectively. I also say an SMTPPORT config there, but
wasn't sure if these could be overridden by site/list mm_cfg.py files.
> 
> >  I don't know where I got the idea that it delivered outgoing mail on its
> >  own. So yes, nothing to fix, just set up the proper outgoing queue runner
> >  with a few sendmail configs.
> 
> 	Not correct.  The easiest way would be to run two copies of the 
> MTA on your machine.  One copy listens on port 25, and runs normally. 
> This is for inbound mail.

> 
> 	The other copy of sendmail listens on a different port (maybe 
> port 26) on the "localhost" interface only, and on the command-line 
> you give it a different configuration file to work from -- that 
> configuration file causes the second copy of sendmail to work in 
> "queueonly" mode, does not have any queue runners, and has a 
> different mqueue directory.  You then control the queue runners for 
> that second instance of sendmail yourself.
> 
> 	You then configure Mailman to deliver all outgoing mail to this 
> second copy of sendmail.
> 
I'm assuming you mean separate sendmail.cf files here when you say 'copy'. I'd
have to think how the second one would differ, and that just running instances
of sendmail configured on the command line wouldn't cut it (much like
MailScanner does, if you're familiar with it's startup scripts).

> 
> 	I still think that this is not a good configuration for doing 
> throttling, because you have only indirect controls over the number 
> of message deliveries that occur in a given unit of time.
> 
> 	The milter technique would be a much more direct method of 
> control over throttling.

OK, I guess I need to sit down and try out these both. I'm going to set up a
test box tomorrow, and see how both works. Of course, I can't simulate the load
it should generate, but for now, I just want to make sure I can do it, period. I
know I can get the milter working, but will have to play with the thousands of
config options for sendmail.
> 
> >>  It's an issue that should be easy enough to fix through modifying
> >>  the configuration file.

Easy for you to say. I have read the Bat Book 2e scantilly, but you know how it
is if you don't need something right away - you usually forget it.

The test box will be running 8.13, which is new here, so a few new options.


> >
> >  I hope you mean sendmail configuration file!!!
> 
> 	Yes.

I sure didn't get much right on the last post, huh?

Thanks very much. You are much too kind and patient.

Steve
> 
> -- 
> Brad Knowles, <brad at stop.mail-abuse.org>
> 
> "Those who would give up essential Liberty, to purchase a little
> temporary Safety, deserve neither Liberty nor Safety."
> 
>      -- Benjamin Franklin (1706-1790), reply of the Pennsylvania
>      Assembly to the Governor, November 11, 1755
> 
>   LOPSA member since December 2005.  See <http://www.lopsa.org/>.
> 




-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/