[TriLUG] Another seal not yet broken... using Greylisting

[jonc]

Thanks for the heads up Tanner.

I vetted the script before I put it on my server. It does exactly what I
wanted it to do. Also, it also came with Wietse Venema's stamp of
approval - which is (in my opinion) a pretty damn good endorsement.

There were a couple of others that I also considered, but since I wanted
to get it up and tested this weekend I didn't have time to vet those.

<jon quick checks the latest load stats again...>

Yep, the load stats this morning are much lower than normal :-)
The greylisting must be slowing down those evil bots!

There is a greylist app written in C (described on one of the excellent
sites you point to below) that I'm going to look at later on, but for
now this simple perl script is a champion.

11 hours + after greylisting was initiated and no spam has made it
through. Very impressive. 
 
On Mon, 2007-01-29 at 10:34, Tanner Lovelace wrote:
> On 29 Jan 2007 01:26:53 -0500, Jon Carnes <jonc at nc.rr.com> wrote:
> > Thanks for the great discussion.
> >
> > I just finished my last tests on my mail server - which now does
> > Greylisting! I'm using a simple perl script and a few added lines to my
> > Postfix configurations. All in all a fairly simple addition to my other
> > anti-spam tools.
> >
> >http://www.opensource.apple.com/darwinsource/Current/postfix-144/postfix/examples/smtpd-policy/greylist.pl
> 
> Um, Jon, greylisting... great!  But, I'm not so sure I'd trust it to
> a random example script.  I'd very much recommend using either one
> of postgrey[1] or gld[2] (grey listing daemon).  I've heard of major performance
> problems with implementing greylisting in a hap-hazard way and I know both
> of those daemons have worked hard to solve problems like that (and a random
> script off a website almost certainly hasn't).
> 
> Cheers,
> Tanner
> 
> [1] http://isg.ee.ethz.ch/tools/postgrey/
> [2] http://www.gasmi.net/gld.html
> 
> 
> 
> -- 
> Tanner Lovelace
> clubjuggler at gmail dot com
> http://wtl.wayfarer.org/
> (fieldless) In fess two roundels in pale, a billet fesswise and an
> increscent, all sable.