[Mailman-Developers] chunkify suggestion, with patch.

[J C Lawrence]

On Mon, 17 Jul 2000 17:52:36 -0400 (EDT) 
Barry A Warsaw <bwarsaw@beopen.com> wrote:

>>>>>> "DF" == Darrell Fuhriman <darrell@grumblesmurf.net> writes:

DF> Here's a new patch for SMTPDirect.py.  Who decides what get
DF> included in the release, BTW?

> I don't think there's time to get this into 2.0 final.  I'd like
> to get more feedback from people on whether this works and is a
> good idea.  My own intuition says that Mailman should do as little
> work as possible here.  I think a good MTA can always do a better
> job than Mailman at deciding the best delivery strategy.  Ideally,
> you'd pass the whole recip list off to the MTA in one huge chunk.

While I agree in principle, software isn't always so generous.  If
you look at the Huge Lists thread (that was the Subject wasn't it?)
we hit this aspect from several directions.

My conclusion:

  Yes, don't try to outguess or outsmart the MTA.  Its not worth it.
That's the MTA's job and it will (always) do it better than you.
That said, you *can* help it.  Domain sorting and clever RCTP TO
bundling will help many MTA's do a better job.  It won't help great
MTAs, but it also won't hinder them or make them slower.

More simply, this patch should make Mailman scale better on marginal
systems.  

> I'm less inclined to coddle inefficient MTAs.  Still, this may be
> a good idea if it's properly configurable.

Inefficient is relative.  The (potential) reduction in spool
contention during deliveries (fewer locked and over-lapping spool
entries) could be quite a help to Exim.  I don't know Postfixes
spool handling well, but I belive it would help there too.

-- 
J C Lawrence                                 Home: claw@kanga.nu
---------(*)                               Other: coder@kanga.nu
http://www.kanga/nu/~claw/        Keys etc: finger claw@kanga.nu
--=| A man is as sane as he is dangerous to his environment |=--