(i'll describe my situation first, the questions are at the end) I managed to make SpamAssassin talk to my Postfix MTA. Brief description of the method:
1. Get & install SpamAssassin (obvious) 2. Get spampd http://www.worlddesign.com/index.cfm/rd/mta/spampd.htm Install various Perl modules required by spampd 3. Run spampd: spampd --port=10025 --relayhost=127.0.0.1:10026 4. Configure Postfix to route messages through spampd: [main.cf] content_filter = smtp:[localhost]:10025 [master.cf] localhost:10026 inet n - n - - smtpd -o content_filter= -o local_recipient_maps= -o myhostname=localhost And then just reload Postfix and that's it. I'm using spampd because i believe it's better to send messages between daemons over TCP, instead of fork()/exec()-ing external helpers (like spamc). I discovered spamproxyd only afterwards. My questions are: Has anyone compared spampd and spamproxyd in terms of speed and reliability under high load? What is the best solution for interfacing Postfix and SpamAssassin for servers under high load? (seems to me like spampd/spamproxyd are the best idea, but perhaps i'm wrong) I sort of like spampd, it has a pre-fork model "a la Apache", and some other nifty things, but i wonder if there are any possible improvements to it... Did you considered including spampd with SpamAssassin, or do you think spamproxyd fits the bill? -- Florin Andrei http://florin.myip.org/ ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
