Scanning for spam

For our colocated server, I'm working on integrating SpamAssassin with the MTA on the server, Postfix. The latest SpamAssassin has features I really like, but I seem unable to correctly implement it the way I want. The easy setup is site-wide, with common settings for every user. But the nice thing about the latest version is that one can use an SQL database (PostgreSQL in our case) with personal settings for each email address. I wish it was as simple as it sounds, I'm unable to make it work the way I want.

I'm trying to use SpamPD to act as a transparant smtp proxy between Postfix and Amavisd-new. I could use Amavis to invoke SpamAssassin, but Amavis doesn't support all the options SpamAssassin has in a transparant way and I actually find it a bit of a sloppy implementation. I mean, it works kinda nice, but it doesn't really use the Unix way-of-life, all kinds of small, independent programs with their own specialty. When SpamAssassin gets new features, Amavis can't use them. So I never really worked with this setup. The trouble with spampd seems to be that it doesn't use the SQL setup as I have told it to. The normal spamd, the one that doesn't work as a smtp-proxy but is distributed with SpamAssassin, can use it, though, with special options that spampd doesn't support. All mail scanned by spampd is processed as user 'root', which is not something I want...

Bart just suggested I should take a look at DSpam, which actually has all the features I want and acts like a transparant SMTP proxy, just the way I like it. It even includes a user interface for each seperate user. I think. I'll take a better look at it tonight.


Comments powered by Disqus