Brian, what strikes me is that bayes is that - you don't have any network tests - do you run spamd/spamassassin with the '-L'-Option? - bayes doesn't show up either in the results but there seem to be tries to autolearn. My guess would be to check permissions in /etc/mail/ - does the user running spamassassin have the rights to actually access his database???
just my two euro-cent Dirk > If I use spamassassin -D --lint then it reveals that I'm at 3.0.2 > > I have posted the x-spam-status from 15 messages at > http://www.meehanontheweb.com/xspamstatus.txt > (the "software_spam_rule", which looks for 'software' in the subject, is > one I wrote in local.cf) > > Autolearn sometimes says "failed" but most often says "no". > > Sans rules, here is what I have in local.cf: > > rewrite_header Subject ***SPAM(_SCORE_)*** > > dns_available yes > required_score 4.0 > bayes_path /etc/mail/spamassassin/bayes > use_bayes 1 > bayes_auto_learn 1 > bayes_file_mode 0777 > report_safe 0 > bayes_ignore_header X-purgate > bayes_ignore_header X-purgate-ID > bayes_ignore_header X-purgate-Ad > bayes_ignore_header X-GMX-Antispam > bayes_ignore_header X-Antispam > bayes_ignore_header X-Spamcount > bayes_ignore_header X-Spamsensitivity > > > > > On Sat, February 4, 2006 23:50, Matt Kettler wrote: >> Brian S. Meehan wrote: >>> My question is, why is it only catching 49% of spam >>> messages? I have the required # set to 4.0 >> >> That's pretty low.. Some questions: >> >> 1) What version of SA are you using? >> >> 2) can you post an X-Spam-Status header from one of your spams that >> didn't get caught? >> >>> Are there things I can put in place, other rules that are preformed or >>> something to catch more spam? >> Yes, but a hit rate that poor suggests there are other problems to look >> for. In particular, check for spam that matches ALL_TRUSTED. >> >> If you ever get any spam (or any external email) matching ALL_TRUSTED, >> please read this for a fix: >> >> http://wiki.apache.org/spamassassin/TrustPath >> >>> Autolearn always seems to be off when I look >>> at the headers of spam messages, caught and uncaught. >> Off? what exactly do you mean by autolearning is off? Do you mean >> autolearn= "no" ,"disabled" ,"unavailable", or "failed"? Each of these >> has different implications as to why autolearning did not occur. >> >> >> > > >
