Re: [Mimedefang] spam score different from when scanning via mimedefang

2014-10-17 Thread info
A quick update on this guys, i have added a DNS caching server, added KAM.cf. Spam did get better, however bayesan still sucks in my configuration. here is an email i got 15 min ago, scores like this: -Spam-Score: 1.308 (*)

Re: [Mimedefang] spam score different from when scanning via mimedefang

2014-10-17 Thread Steffen Kaiser
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Fri, 17 Oct 2014, i...@bsolution.net wrote: Spam did get better, however bayesan still sucks in my configuration. here is an email i got 15 min ago, scores like this: well, Bayes could have learnt in that time. However, did you checked if

Re: [Mimedefang] spam score different from when scanning via mimedefang

2014-10-17 Thread info
On 2014-10-17 10:08, Steffen Kaiser wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 well, Bayes could have learnt in that time. However, did you checked if the Bayes database has more than one user? Maybe SpamAssassin uses two users internally. Thanks Steffen i will look into that.

[Mimedefang] spam score different from when scanning via mimedefang

2014-10-16 Thread info
Hello guys, i am desperate. this question has been asked many times - yet the resolution does not apply to me because most cases mimedefang runs a different user. I also Searched many places, read a lot of docs and posts - still can't solve the mystrey. i have a relatively straight forward

Re: [Mimedefang] spam score different from when scanning via mimedefang

2014-10-16 Thread Kevin A. McGrail
On 10/16/2014 11:26 AM, i...@bsolution.net wrote: Hello guys, i am desperate. this question has been asked many times - yet the resolution does not apply to me because most cases mimedefang runs a different user. I also Searched many places, read a lot of docs and posts - still can't solve

Re: [Mimedefang] spam score different from when scanning via mimedefang

2014-10-16 Thread info
On 2014-10-16 12:13, Kevin A. McGrail wrote: I don't know that you are missing anything. I'm assuming time has elapsed between these two tests because you are showing different Bayes scores, different RBL hits, etc. which would indicate that the RBLs reactively added information after you