Keith,
   Look through the archives.  Somebody posted an answer to this before. 
  Basically, what they did, was follow the "simple" configuration 
described in the README.filter and in the shell script they used 
"spamc".  I don't know that anyone has used spamproxyd, but the trick is 
that you must bring mail back into Postfix on a DIFFERENT port than the 
default 25, and in the configuration you must set a variable which tells 
postfix not to filter the message again.  This is detailed in the 
README.filter as the "Advanced" configuration (sorry, don't have access 
to the docs at the moment).
   So, you should be doing something like this:
   Postfix: Listening on Port 25 and 10025
   spamproxyd: Listening on Port 20025 and configured to forward mail to 
localhost:10025.
   (NOTE: Don't forget to block ports 10025 and 20025 with your 
firewalling rules!  We don't want you to create an openrelay by bouncing 
mail off spamproxyd into postfix!)
   In the postfix master config, setup port 10025 for smtpd and add a 
variable such as "NoFilter=yes" (dont' remember exactly what you should 
use, check the README.filter).
   Mail flows as follows:
MUA-->Postfix(25)-->spamproxyd(20025)-->Postfix(10025)-->Destination(smtpd 
or local delivery).
   One thing to note is that, if this is your default internal mail 
server, the outgoing messages will ALSO be filtered.  This may not be 
desirable :^).  I suppose you could add some sort of check at the 
beginning of spamproxyd that would check if the first "From " in the 
message is a local machine and if so, pass the message back to postfix 
unchecked ... you'll have to experiment with that.
   This is something I'm planning to do eventually, so let me know how 
it goes ;^).
   One of the really cool things you could do is add the ability to 
delete/bounce messages that are marked as spam in spamproxyd.  Since 
you're using MySQL, you could simply look in the spamassassin tables for 
a configuration like (spamstatus=[deliver|drop|bounce]) and add that 
option to the webpage you use to let users manage SpamAssassin on their 
accounts.
   I'm doing something similar by adding ":DELETE" to the "****SPAM****" 
subject tag if the user wants all spam dropped.  The procmail recipe 
that handles delivery looks for "****SPAM:DELETE****" and delivers it to 
/dev/null if it's there (well, actually it delivers it to a folder I 
review for false positives periodically ... get's about 300 spams a day 
with about 35 users, and growing rapidly ;^).
   Hope this is helpful.
   Don

Keith Pitcher wrote:

> Anyone have spamassassin working with postfix with the mysql patch
> making everyone a virtual user?
> 
> I'd like to run everything through spamproxyd, the directions appear
> simple enough. However mail just loops constantly in every configuation
> I've tried.  Postfix -> spamproxyd -> postfix -> spamproxyd ....    
> Ian was nice enough to give me his example config even, to no avail. 
> 
> I'm thinking that since the mysql/virtual section of postfix is 1/3 of
> the code of the "normal" postfix delivery, things just aren't able to
> work with filters right. 
> 
> Could someone who has this working please prove me wrong so I can get to
> using spamassassin again?
> 
> Thanks,
> 
> Keith
> 
> 
> 
> _______________________________________________
> Spamassassin-talk mailing list
> [EMAIL PROTECTED]
> https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
> 
> 
> 



-- 
--------------------------------------------------------
Donald L. Greer, Jr                  [EMAIL PROTECTED]
System Administrator                 Voice: 512-300-0176
AustinTX                        http://www.AustinTX.COM/
   All opinions are my own.  Flame me directly.

"I don't necessarily believe software should be free...
but if you pay for it, it should work!" -- Me


_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to