John Schmerold wrote:
> That, of course is one of the reasons I wrote, I suspect we could do
> better. An obvious area for our situation is to deny based on
> recipient email address (if we don't exist, why are we accepting mail
> for them).
> 

That IS a big one, yes...

> This box is a filter for several clients using mail servers that
> include Exchange, cPanel, Groupwise & Vmailmgr, so configuring
> recipient addresses is easier said than done.
>

But well worth it...

We use a single DB instead of a complex verify-mode router-walk.

Arguably less efficient to do an SQL call than walk Exim routers, BUT - there 
is 
just ONE place to look or 'maintain' (even 'postmaster' has an entry there, as 
do our valid MLM input addresess).

..and there are *many* ways to grab the data from all manner of disparate input 
sources and formats and suck it into the DB.

> We're using several (probably too many) rbls, that helps, guess
> greylisting is next.
>

BT, DT, GGTS.

But a simple 15-second delay whenever a 'fault' is hit (else NOT) has been just 
as effective, and with no discernable delay on decent traffic..

Seems a goodly number of 'modern' zombots are programmed to not waste 'their' 
time on recalitrant victims....

W/R rDNS fail, HELO mismatch, RBL 'hit', header format and mime faults - 
generally too much 'falsing' if used as pure go no-go.

Instead, you can give these each weighted 'demerit' scores based of severity, 
carry them about in bespoke and/or cumulative acl_c(n), transfer to acl_m(n), 
do 
gt/lt/eq testing against system-wide, per-domain, or per-user tolerance levels.

HTH,

Bill


-- 
## List details at http://lists.exim.org/mailman/listinfo/exim-users 
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Reply via email to