Re: SMTP-Interface for pyzor

[Guido <lis...@gu...>]

Thanks for your reply.

On (09-06-26 00:13), Benny Pedersen wrote:
> On Thu, June 25, 2009 19:12, Guido wrote:
> > As Amavis/SpamAssassin is very resource hungry I would like to reject
> > mails if pyzor returns a number of reports of lets say 20.
> pyzord is imho very fuzzy in there sigs so use it as a mta reject is imho a bit dangerous
> > Of course the mails should be rejected before entering Amavis and SpamAssassin.
> no problem for me in this way
Sorry, but I can't understand what you mean. I mean rejected by Pyzor.
How can this work? As far as I can understand this, your statements are
conflicting? :-)

> >  - Therefore I am wondering if there is any solution to use pyzor
> >    as an  SMTP proxy. Functionality like dccifd[1] provides for DCC[2].
> i would like it to be used as a spamtrapper, learn all as spam if sent to newer existsing recipient
> or extend it with spf in the proxy mode if that would be coded, or with spf checking ?, learn as whitelist if spf pass ?, all else
> learn as spam ?
Currently our Pyzor server learns mails from a specific mailbox. But we
plan to set up a spam trap server. I think it should work by simply
adding an entry to /etc/aliases like this one:
catchall: |"/usr/bin/pyzor report"
Where the server accepts all mails for a spam trap domain.

Sure, a dedicated pyzor MTA may be faster doing this job.

> > Are there any other suggestions how to use pyzor as a before queue
> > filter? On the homepage I could only find how to call pyzor by
> > SpamAssassin or by procmail.
> 
> procmail is fine for when recipient is not virtual user

Procmail does not work for us because the mail server must relay to
other servers. Beside that procmail usually comes after
Amavis/SpamAssassin.

Guido