Date: Thu, 12 Jan 2017 15:16:22 +0000 From: RW <rwmaillists@googlemail.com> To: freebsd-questions@freebsd.org Subject: Re: spamassassin not lethal anymore Message-ID: <20170112151622.40493b27@gumby.homeunix.com> In-Reply-To: <2463a238-e10f-e81d-cab1-5a7eaf774590@pinyon.org> References: <2463a238-e10f-e81d-cab1-5a7eaf774590@pinyon.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 11 Jan 2017 13:45:47 -0700 Russell L. Carter wrote: > Howdy, > > I've been happy using postfix+spamassassin for a long long > time, and it's always worked great. However in the last > few weeks it's not been scoring spam high enough, and in the > last 10 days the spam is getting through in a torrent. I > see a lot of scores in the 1-2 range, for what is obviously > spam. I'm not really comfortable setting the threshold to > 1, say. It sounds like you are either not using Bayes or it's not working well. If you use Bayes with auto-training it is liable to go "off the rails". If you're just filtering your own mail the best thing to do is to turn-on Bayes, turn-off auto-training, and train manually (running sa-learn as the correct unix user). Another thing is that if you do spam filtering, you should run your own non-forwarding DNS resolver. Shared caches can get blacklisted by domain and IP blocklists because they can go over the usage limit for a single IP address.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20170112151622.40493b27>