Hi Jeff.
Am Montag, 24. September 2007 schrieb Jeff Jansen:
I'm sure many of us would appreciate it if you'd update the list in the
future on how well this works at preventing spam. (At least I know I
would.) I've seen mixed reviews about the efficacy of greylisting, and
I'd love to hear first-hand from someone using courier how effective it
really was.
For some reasons, I wrote a greylisting-filter from scratch instead of using
comeagain.py or some variants. But I think the efficiency would be
comparable.
After a month of usage (must have been in february this year), I wrote this
(german) article in our company blog:
http://schokokeks.org/ein_monat_greylisting
For non-germans, here the facts (counters are of a 24-hour-period):
* accepted messages: 2471
* auth-smtp-messages (unfiltered pass-through): 981
* temporary rejected messages: 3554
* temporary rejected but accepted later ("second try"): 273
* messages that didn't come again: 3281
* overall arriving messages: 6025
That means:
Only 273 out of 3554 delayed messages came again.
So less than 10 % of the messages with unknown receiver/sender- or ip-adresses
were accepted at all.
and
We rejected 3281 messages out of 6025 arriving mesages due to greylisting.
That's more than half of all messages.
cu, Bernd
15 messages in net.sourceforge.lists.courier-usersRe: [courier-users] Greylisting with ...
