Qmail with Spamassassin (bogofilter to be installed) Question

[Ryan Pagquil]

At 04:14 PM 3/24/2006, David Relson wrote:
>On Wed, 22 Mar 2006 17:12:21 -0800
>Ryan Pagquil wrote:
>
> > Hi Bogofilter gurus,
> >
> >       I'm using qmail and spamassassin together. Now that we receives a
> > lot of spam mails I want to use bogfilter to minimize it or if lucky
> > zero it. This is the scenario.
> >
> > I'm running two MX servers MX1 and MX2, both runs qmail and
> > spamassassin. All mails bound for our domain will be forwarded to
> > another mail server which serves to be the inner mail server runs
> > qmail but no spamassassin. Now I want to use bogofilter for this
> > setup but I don't know how to start.
> >
> > Here are my questions.
> > 1. Does bogofilter should be installed in the MX servers or in the
> > inner mail server?
> > 2. How will I get ham mails? Usually users only reports spam mails?
> > 3. Whos responsibility to train bogofilter, users?
> > 4. In training bogofilter, should ham mails = spam mails?
> >
> > Please help...
> >
> > Thanks in advance,
> > Ryan
>
>Hello Ryan,
>
>You can install bogofilter either way (on the 2 outer servers or the 1
>inner server).  It's probably easier to use the inner server as you
>won't have to worry about synchronizing wordlists.
>
>For initial training, it works well to feed bogofilter with 1,000 or so
>ham and an equal number of spam.  After that, you should have your
>users forward any classification errors to you so that yo_ can use
>them to train bogofilter.  Alternatively, with bogofilter's tri-state
>classification, you (as the administrator) could have procmail save any
>messages classified as "unsure" and then you could make the spam/ham
>determination and use them to train bogofilter.
>
>Alternatively if you trust your users, you could have them forward
>incorrectly classified messages to addresses like
>"this.is.spam at yourdomain.com" and "this.is.not.spam at yourdomain.com" and
>have a script to feed them to bogofilter.
>
>HTH,
>
>David


Hi David,

I need to feed bogofilter ham mails and it means that I need to get 
valid mails from my users and feed it to bogofilter? My user's emails 
directly goes to ~theirhome/Maildir/new or ~theirhome/Maildir/cur so 
I need to get those for my 1000+ users? Is there any simplified 
approach for this? No problem for spam mails since they are reporting it.

Thanks,
Ryan