hello,

In the past I posted a question about how to create a mail-server, and for the moment I created a working smtp linux server (postfix)... and it works !! :-)

but what i don't understand how to configure a pop3-server for several windows clients ?

I read information about fetchmail, courier-imap, qpopper, ... but I still not understand how to configure it like I wish?

I want that my linuxserver receive all the pop3 mails from my ISP(telenet, in.pandora.be), and it also checks for spam with SpamAssasin. and that then with a windows client (outlook) will connect to my linux mailserver and get all the mails checked !

thanks, in advance,

balki

Install fetchmail as a system-wide service. It will fetch e-mail from external servers (ISPs) using POP3 protocol. Check fetchmail docs on how to configure it to run as a service and how to set up individual user mapping (usernames and passwords on ISP's pop3 accounts -> usernames on your server).
Fetchmail will pass the e-mail to procmail.
Configure procmail to pass the e-mail to spamassassin. This is the standard way spamassassin is supposed to work, so I guess you won't have any problems here.
Then you need a pop3 server on your server. It will server e-mail to the windows clients using POP3 protocol. No special needs to configure this one - should be fine with default config.

ok,

thanks !! but I have also another question? how can I configure it like a email-relay? I mean, how can I give all the mails from my linux ( checked by spamassassin) to Exchange Server ? MTA --> MTA ?

thanks,

Jonay

Have your linux MTA relay (forward) all mail to the Exchange MTA.

How you configure this depends on which linux MTA you use.

Sincerely
Bill Jennings

Take a look at the postfix documentation regarding the "transport" table. i.e.
This is what I do. I frontend my exchange server with postfix; which also calls spamassassin prior to relaying to the exchange server.

NOTE: Prior to making this simple transport change, be sure you configure postfix to vaildate the recipients on your exchange server prior to relaying. Otherwise, an e-mail addressed to bogus@mydomain.com will be forwared to your exchange server; which will then issue a bounce back to a non-existant address. In postfix, this is done by configuring the "relay_recipient_maps". FWIW: I configure postfix to issue an LDAP request to my exchange server to validate the recipient e-mail address (mailbox), but you can also simply create a file with a list of valid mailbox's on your exchange server. See the postfix documentation regarding the relay_recipient_maps.

Long pause...

OK, I just re-read this thread. If your using fetchmail, then you would not have the problem I describe above. I was referring to a live MTA that receives e-mail directly from the internet, not piped through fetchmail. But I still "think" you may need to set the "relay_recipient_maps" in postfix before it will relay to your exchange server using the transport maps. Unfortunately, I have no way of testing this at the moment.