On Wed, Dec 20, 2000 at 02:00:08AM +0000, Mark Delany wrote:
> > Reason being, anyone can point their domain's MX records to your server
> > and then freely send mail to/through it.

Nope ;-)
If I would implement it I'd have a small file e.g. "mxips" which would
contain IPs/IP ranges and only if a MX hosts IP is in there I'd accept it.

> > I'm curious to know if qmail starts having performance issues at higher
> > numbers of rcpthosts. We're currently running about 1800 domains, most of
> > them in the morercpthosts file.
> 
> cdb has pretty linear performance regardless of the number of
> entries. cdb is used for databases that can typically get *much*
> larger, such as name server data and /var/qmail/users. 1800 entries is
> nothing for cdb.

This is true.
However one of our mailservers is sitting on an elder machine with only
one disk (we're moving it to a newer/fast machine with raid by now)
and what I did is experience the last weeks was a performance loss
simply by having qmail-smtpd read in about 700 MB rcpthosts and
qmail-remote read in a 300 KB smtproutes file on every incarnation.
a cdb would do any good here, as the lookup would be faster, but a cdb
is even bigger than the plain files.
What really helped in that situation was a small mod to qmail-smtpd
(1.01) to only read in rcpthosts if relayclient wasn't set. So the
problem is (soon was ;-) IO bound not CPU bound.

What I'd also wanted to overcome is the "explanative" overhead of
managing rcpthosts for customers that order/release domains without
telling us and then complain that mail gets bounced.
Argueing them is IMHO a waste of time ... they're customers, they don't
want to think oder understand how things work, they want it to work and
go mad on you if it doesn't regardless whose fault it is.

        \Maex

-- 
SpaceNet AG               |   http://www.Space.Net/   | Stress is when you wake
Research & Development    | mailto:[EMAIL PROTECTED] | up screaming and you
Joseph-Dollinger-Bogen 14 |  Tel: +49 (89) 32356-0    | realize you haven't
D-80807 Muenchen          |  Fax: +49 (89) 32356-299  | fallen asleep yet.

Reply via email to