> -----Original Message-----
> 
> There are some big names that play badly with greylisting. They play
> badly with greet-pause, too. A problem I've seen with 
> greylisting is the
> round-robin MTA pool. Each is told in turn to come back later 
> and if the
> pool is large it can take a long time to cycle through all of 
> them. You
> have to be careful how you screen the addresses.
> 
> dp

The greylisting scheme I have implemented works at the DATA phase.  It
uses the sender IP address (top 24 bits only), the sender e-mail address
and header date field to form the key for the message.  Once a message
has passed the greylist test the original sender IP address (full 32
bits) is placed in a whitelist.

So, a particular server only needs to demonstrate once that it re-tries
and will then be let through in future.  By using the top 24 bits of the
IP address in the key I hope to cope with a message being re-tried by a
different MTA.  I have not encountered such a problem yet.

I have had a couple of instances where there was a problem because
people had written their own code on web servers.  They did not re-send
the same message, but re-generated it when re-trying and so gave it a
new date header.  In both cases they modified their code when I
explained the problem.

Phil.
--------------------
Phil Chambers
Postmaster
University of Exeter

_______________________________________________
Help us build a comprehensive ClamAV guide: visit http://wiki.clamav.net
http://www.clamav.net/support/ml

Reply via email to