How do you count e-mail?

[Alessandro Dentella]

> Pflogsumm is a bit confusing when used with MailScanner, since we HOLD
> everything we accept, sort of.

 sure, we can just disreguard that information

> So your "ratio" would be the comparision (more or less) between ~1.1
> million message "attempts" and the actual delivered messages.
> Paints a pretty picture;-).

  amaizing ratio, to me. But I think that the 1.1 million attempts is the
  base to calculate how many times I will make dns request, to that that's
  probably the number to compare with 250.000 mail/day suggester as a
  turning point by rbldnsd website.


> 
> Just out of curiosity... what are the top rejections?

message reject detail
---------------------
  RCPT
    Helo command rejected: need fully-qualified hostname (total: 536368)
        7235   telesp.net.br
        5312   ono.com
        4455   vtr.net
        2990   veloxzone.com.br
        2814   89.104.106.98
        2698   211.173.152.57
        2696   ntl.com
        2480   ppp91-76-9-202.pppoe.mtu-net.ru
        2347   91.146.59.228
        2337   211.179.104.107
        2292   211.228.99.206
        2286   122.43.109.217
        2272   fibertel.com.ar
        2208   qualitynet.net
        2190   brasiltelecom.net.br
        2158   ctbcnetsuper.com.br
        2120   volia.net
        2103   211.105.59.248
        2065   asianet.co.th
        2061   blueyonder.co.uk


> >    2. which is the average % spam that is 'fisiological' to accept in a fine> >       tuned server?

> Sorry, don't really know what you mean with "fisiological"...

 spam that even if I finely tune my antispam machinery I'll never stop.
 One for all. I had for some weeks some very low control but I had
 autowitelist on. Bayes started giving negative score to spam. Just
 switching bayes off I could get mush better filters. 

sandro
*:-)