Carlos E. R. wrote: > > The Friday 2006-11-17 at 10:29 -0800, J Sloan wrote: > >> Hi Chuck, one mistake I see in the script is that it starts both amavisd >> and spamd. If you're running amavis, you should not start spamd, since >> it's a duplication and a waste of resources at best. amavis essentially >> takes the place of spamd. > > That's generally not true. > > You can, for instance, disable spam checking in amavis, and use spamc > later, with per-user bayesian databases, which is a more powerful > configuration than amavis. In fact, I don't know how to use per-user > bayes databases with amavis-new: it uses a global, system-wide, database.
In general, spam is spam. In the odd case that someone wants spam, there are mechanisms within amavis to allow a user increased spam, up to allowing all spam to that user. While in some cases I suppose one might use amavis for virus only and use spamc/spamd for spam detection, that is the exception, and I've never seen anybody purposely use both spamd and amavisd to run spam assassin. Loading the sa perl classes via amavis is very efficient. For any serious spam fighting scenario though, we run maia mailguard, which uses a specially modified version of amavisd-new. In mailguard, the per user white/blacklists, and SA thresholds can be set by each user for himself in the web interface. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
