MailScanner and Spamassassin

[Kevin Miller]

Yeah, you definitely should.  Lots of training bound up in the old database.  Also, IIRC, bayes doesn't kick in until it sees 200 hams and 200 spams, so there's a period where it's learning but not contributing.  Or something along those lines...

 ...Kevin
--
Kevin Miller
Network/email Administrator, CBJ MIS Dept.
155 South Seward Street
Juneau, Alaska 99801
Phone: (907) 586-0242, Fax: (907) 586-4500
Registered Linux User No: 307357
From: mailscanner-bounces at lists.mailscanner.info [mailto:mailscanner-bounces at lists.mailscanner.info] On Behalf Of Mike Watson
Sent: Friday, October 19, 2012 2:03 PM
To: MailScanner discussion
Subject: Re: MailScanner and Spamassassin

New box I didn't copy the database but it's an idea.

Thanks,  mw


--



"Lose not thy airspeed, lest the ground rises up and smites thee."

  -- William Kershner

http://crucis-court.com

http://www.crucis.net/1632search

On 10/19/2012 03:17 PM, Kevin Miller wrote:
My understanding is that you use sa-learn to train the database.  If you get a false positive you feed it to spamassassin as non-spam.  If spam slips through, you feed it back to spamassassin as spam.  I expect sa-learn is also called during the scanning process in the background.   I don't think you need to do anything special w/your users unless a message is mis-tagged.  The previous bayes database should be still working and being updated...

 ...Kevin
--
Kevin Miller
Network/email Administrator, CBJ MIS Dept.
155 South Seward Street
Juneau, Alaska 99801
Phone: (907) 586-0242, Fax: (907) 586-4500
Registered Linux User No: 307357
From: mailscanner-bounces at lists.mailscanner.info<mailto:mailscanner-bounces at lists.mailscanner.info> [mailto:mailscanner-bounces at lists.mailscanner.info] On Behalf Of Mike Watson
Sent: Friday, October 19, 2012 11:49 AM
To: MailScanner discussion
Subject: MailScanner and Spamassassin

I suppose this is more a Spamassassin question than MailScanner but I'll ask anyway.

I've just finished setting up a mailserver using CentOS 6.3, Dovecot, sendmail, and MailScanner-Spamassassin. This is an upgrade from an older linux server that also ran an older version of MailScanner-Spamassassin.

I know SA runs sa-update via cron. That's working. But...is it necessary to run sa-learn for each user as well to detect SPAM that gets through MailScanner?  There has to be a reason for sa-learn or it wouldn't be included in the SA package. I do have Baynes activated for SA.

Is there any reason to run sa-learn on the spam files?

mw



--

--



"Lose not thy airspeed, lest the ground rises up and smites thee."

  -- William Kershner

http://crucis-court.com

http://www.crucis.net/1632search


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.mailscanner.info/pipermail/mailscanner/attachments/20121019/3e76e4e4/attachment.html