我有很多垃圾邮件的问题。 我的服务器设置是VPS与CentOS 6.5和Dovecot + Postfix + Spamassassin。
这里是其中一条消息的来源:
Return-Path: <[email protected]> X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on xxxxxx.xx X-Spam-Level: X-Spam-Status: No, score=-0.5 required=3.0 tests=HTML_MESSAGE, HTML_MIME_NO_HTML_TAG,MIME_HTML_ONLY,RP_MATCHES_RCVD,URIBL_BLOCKED autolearn=ham version=3.3.1 Delivered-To: [email protected] Received: by mx.xxxxxx.xx (Postfix, from userid 5001) id 584194399E; Sun, 14 Sep 2014 04:34:03 +0200 (CEST) Received: from tagbook.co.nz (tagbook.co.nz [119.47.112.73]) by mx.xxxxxx.xx (Postfix) with ESMTP id 6089C43119 for <[email protected]>; Sun, 14 Sep 2014 04:33:59 +0200 (CEST) Date: Sun, 14 Sep 2014 14:33:56 +1200 From: "Nancy Mitchell" <[email protected]> Reply-To:"Nancy Mitchell" <[email protected]> Message-ID: <[email protected]> To: [email protected] Subject: RE: Lol adooya X-Priority: 3 (Normal) MIME-Version: 1.0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: 8bit <h2><a href="http://www.dirtymath.co.kr/data/question/g0702/login.html?ZGpwQmRqcGxndixndw==">adooya</a></h2> <div> The artifact was instantly forgotten-as I hoped, although it took </div>
正如你所看到的,这个消息只有垃圾邮件得分-0.5。 有人有什么想法,该怎么办? 这很烦人。
根据我上面testing的SAconfiguration,只有RP_MATCHES_RCVD
产生负分(火腿指标)。 RP_MATCHES_RCVD
得分在-1.5和-0.5之间(本地/networking,带有贝叶斯/无贝叶斯)。
您可以考虑closurestesting,如在man Mail::SpamAssassin::Conf
。
〜/ .spamassassin / user_prefs中
score RP_MATCHES_RCVD 0
安装DCC,Pyzor和Razor2。 这显着改善了检测。 如果您仍然收到垃圾邮件,请编写一些元规则(例如DCC +黑名单= +2分)。 我用这种方法每周阻止300多个垃圾邮件,准确率超过99%。