New Guy - Bayes, MS MTRG help please?

[Tim Sailer]

On Fri, Sep 17, 2004 at 10:44:54AM -0700, Greg Deputy wrote:
> A little background.  To be upfront, I'm a novice user with MailScanner
> and all things Linux, so be warned (please be gentle).  I've been a

Welcome to Open Source. "We're Gentle."

> windows administrator and developer for close to 10 years, working
> onsite at MS.  I've recently made the jump to using open source/Linux in

It's about time! Walk towards the light! :)

> my side business hosting domains, email and web apps on a set of windows
> servers sitting in a local datacenter.  The first thing I'm putting into
> production is a MailScanner setup.
>
> The machine is a P3 933mhz, 386Meg ram running Fedora2, Postix,
> Mailscanner, Spamassassin, ClamAV, Razor, MTRG.  When the machine is
> dropped in the datacenter it'll be in front of an Imail server currently
> hosting about 80 domains and 500 email addresses.  The volume is light
> to moderate.  It'll start by doing virus scans on all email, and spam
> filtering for only few domains.

Shouldn't be a problem.

> For the last week I've been running the machine in my home on a cable
> connection, filtering email for a personal domain.  I'm doing that
> mainly to get comfortable with running it, trying out different
> settings, etc.  If something breaks, it only effects me for the moment.

Smart move. I use my home systems for the proving ground, before I
break something that will affect thousands of users. You tend to live
longer that way.

> So far, everything has been going surprisingly well, I should have tuned
> out the anti-open source propaganda much sooner!  There are 2 minor

(no comment, as much as I want to)

> issues I cant seem to find a solution for by searching the faqs,
> archives, google, etc.  I also have a nagging question about postfix and
> MS, below.

[clip mrtg]
I din't run mrtg for Mailscanner, so I'm no help.

> The other issue is Bayes filtering in spam assassin.  I've run about
> 1200 ham and 1700 spam messages through sa-learn, and when I run
> 'spamassassin -D -lint' it says the bayes database is there and happy.
> However, in /var/log/maillog where I have spam logging through
> mailscanner turned on to full so I see all the spamassassin rules that
> trigger, I don't see any bayes rules ever hit.  I see other things,
> including Razor rules in SA after installing razor last night, but still
> no bayes stuff.  I've gone through all the faqs on checking if
> spamassassin is working, its bayes stuff, etc, but cant seem to get it
> to work.  From the command line everything appears like the bayes
> filtering is ok, but it never seems to hit through mailscanner.

Is MailScanner an SA both looking in the same place for the files?
Doublecheck.

> My nagging question is about using postfix and mailscanner.  I've seen
> the posts on forums from the postfix people saying mailscanner is bad,
> DON'T use it with postfix, etc.  I've also seen responses from
> MailScanner people saying that was only the case in earlier versions, it
> works fine now, no problems, etc.  Can anyone give me a final word on
> that issue?  Is it ok?  I'm comfortable with postfix and would like to
> stick with that, but if its not stable I'll get up to speed on sendmail
> and make the switch.

I'm a long-time user of Exim for my commercial stuff (speed under load
issues, let's not start a war over this), but at home I'm using postfix
and MailScanner. The way Julian has written here:
http://www.sng.ecs.soton.ac.uk/mailscanner/install/postfix.shtml

worked straight up for me, not counting the RH-isms, as I run Debian. It's
fast, sensible, and just works.

Tim

--
Tim Sailer <sailer at bnl.gov>
Information and Special Technologies Program
Office of CounterIntelligence
Brookhaven National Laboratory  (631) 344-3001

------------------------ MailScanner list ------------------------
To unsubscribe, email jiscmail at jiscmail.ac.uk with the words:
'leave mailscanner' in the body of the email.
Before posting, read the MAQ (http://www.mailscanner.biz/maq/) and
the archives (http://www.jiscmail.ac.uk/lists/mailscanner.html).