> Hi everyone, > > I want to add Spam Assassin 3.23 to my SPAM counter > measures > It is working into a FreeBSD jails, over Amavisd-new port. > > When I test SpamAssassing via: spamassassin > <sample-spam.txt, > it reacts correctly, raising the GTube signature: > X-Spam-Flag: YES > X-Spam-Checker-Version: SpamAssassin 3.2.3 (2007-08-08) > on ns.levier.org X-Spam-Level: > ************************************************** > X-Spam-Status: Yes, score=1001.2 required=5.0 > tests=AWL,GTUBE,NO_RECEIVED, > > NO_RELAYS,RAZOR2_CF_RANGE_51_100,RAZOR2_CF_RANGE_E4_51_100,RAZOR2_CHECK > autolearn=no version=3.2.3 > > spamd is launched normally as spamd user, and runs > apparently fine. > Amavis does its job pushing email to spamd (filter is > named spamassassin, different from 'spamd' name, > confirming in logs the mail is "checked": > > Dec 26 13:16:39 ns spamd[14210]: spamd: connection from > ns [IP_address] at port 60196 > Dec 26 13:16:39 ns spamd[14210]: spamd: processing message > <[EMAIL PROTECTED]> for spamd:58 > Dec 26 13:16:39 ns spamd[14210]: spamd: clean message > (0.0/5.0) for spamd:58 in 0.0 seconds, 551 bytes. > Dec 26 13:16:39 ns spamd[14210]: spamd: result: . 0 - > scantime=0.0,size=551,user=spamd,uid=58,required_score=5.0,rhost=ns,raddr=IP_address,rport=60196,mid=<[EMAIL > PROTECTED]>,autolearn=ham > Dec 26 13:16:39 ns spamd[14209]: prefork: child states: II > > However, all emails passing though amavis + Spamd (via > port 783) generate a > 0.0 score, even the one with GTUBE signature. > > I dont understand why Spam Assassin test works well and > spamc method totally fails. > > Can someone help? >
Not much help from this, but I wonder how this goes... Normally, standard Amavis does not call spamd at all, but loads it's own copy of spamassassin - as they both run under perl, and amavis is a resident process. No need for spamc/spamd for Amavis. I do run spamd/spamc pair, and have disabled spamassassin in my amavisd-new configuration. I call spamc via maildrop script.
