Dave Koontz writes: > I just noticed that for some reason only some of my messages are > actually being run through Bayesian classifying. I am not sure how long > this has been occuring. I did a google search which did not turn up much > as to what could cause this. > > One suggestion was there was not enough processing threads and/or > connections. I have tried 6 / 200 and it still occurs. > > The other was related to file permissions, but there are no permission > problems and some messages score just fine, and my autolearn is not > complaining about permissions and is working fine. > > I am seeing this in both 3.2.3 and 3.2.4. I even tried creating a new > bayes_toks database to no avail. I even removed 60_shortcircuit.cf > thinking perhaps that was the culprit. > > The only indication is SA's X-Spam-DataBase: Bayes not run. header.
SpamAssassin doesn't add a header like that... what are you using: spamd, MailScanner, amavisd? --j.
