În data de Jo, 06-05-2010 la 21:13 +0200, Stevan Bajić a scris: > On Thu, 06 May 2010 21:34:01 +0300 > Stas Sușcov <[email protected]> wrote: > > > În data de Jo, 06-05-2010 la 19:41 +0200, Stevan Bajić a scris: > > > On Thu, 06 May 2010 18:16:01 +0300 > > > Stas Sușcov <[email protected]> wrote: > > > > > > > Hi, > > > > we were using DSPAM and Postfix for about a year and now we would like > > > > to connect the rest of our MX servers to it. > > > > We got dspam on the main mailserver and we would like to connect a > > > > secondary server to it, so I set dspam to listen to a port, changed > > > > master.cf to use that new port, in short the setup is pretty much the > > > > same I documented here: > > > > https://help.ubuntu.com/community/Postfix/Dspam > > > > > > > > The problem is with the secondary server in master.cf, where I mush tell > > > > Postfix on what port to listen for content from dspam. How do I teach > > > > dspam to respond to the same ip from what it received a new submission? > > > > Secondary server Postfix, receives no response from submitted entries. > > > > > > > Might I quickly repeat what I have understood? > > > > > > Example: > > > MX 1 has IP 1.2.3.4 > > > MX 2 has IP 5.6.7.8 > > > DSPAM has IP 9.7.5.3 > > > > > > DSPAM is behind port 11124 on IP 9.7.5.3 > > > > > > Now you have set in MX 1 to deliver to DSPAM. Something like this: > > > /./ FILTER lmtp:[9.7.5.3]:11124 > > > > > > And on MX 2 you have done the same: > > > /./ FILTER lmtp:[9.7.5.3]:11124 > > > > > > And now you would like DSPAM to deliver back to the IP where the original > > > message came from? Right? In case of MX 1 you would like DSPAM to deliver > > > to 1.2.3.4:10024 and for MX 2 you would like DSPAM to deliver to > > > 5.6.7.8:10024. Right? > > > > > > Actually dspam is on the main server: > > Main server (A): 1.2.3.4 with dspam on port 24 > > Secondary server (B): 5.6.7.8 > > > > A postfix-dspam link: > > /./ FILTER lmtp:[1.2.3.4]:24 > > B postfix-dspam link: > > /./ FILTER lmtp:[1.2.3.4]:24 > > > > I need to get dspam reply submissions accordingly to the source ip: > > DeliveryHost: <getTheSourceIP()> > > > I could code something to allow that. But the source IP is just one part of > the issue. What about the port? To which port should DSPAM connect? > > > > > > > > I found this: > > > > `If you would like to set up DeliveryHost's on a per-domain basis, use > > > > the syntax: DeliveryHost.domain.com 1.2.3.4` > > > > > > > > What's the use-case of such a directive in dspam.conf? > > > > > > > For example MX1 is responsible for the following domains: > > > oracle.com > > > mysql.com > > > postgres.org > > > > > > And for example MX2 is responsible for the following domains: > > > microsoft.com > > > ibm.com > > > novell.com > > > > > > To get DSPAM to deliver messages to the right MX server you would then > > > add something like this to dspam.conf: > > > > > > # Deliver @oracle.com, @mysql.com and @postgresql.org to MX 1 > > > DeliveryHost.oracle.com 1.2.3.4 > > > DeliveryHost.mysql.com 1.2.3.4 > > > DeliveryHost.postgres.org 1.2.3.4 > > > > > > # Deliver @microsoft.com, @ibm.com and @novell.com to MX 2 > > > DeliveryHost.microsoft.com 5.6.7.8 > > > DeliveryHost.ibm.com 5.6.7.8 > > > DeliveryHost.novell.com 5.6.7.8 > > > > > > # Deliver anything else to MX 1 > > > DeliveryHost 1.2.3.4 > > > > > > > So the DeliveryHost per domain basis is not a solution in my case. Both > > servers A and B are managing almost the same domains (with small > > exceptions) and B is also a backup MX for A. > > > This sounds to me like a design issue you have there. I mean if MX B is the > backup for MX A then are you not breaking the purpose of MX B in case MX A is > down? MX A down = no DSPAM available for processing. > > What do you use as storage backend for DSPAM? How do you ensure that MX (A) > and MX (B) know for which users they are responsible? Do you have a central > user/domain management? >
Sorry but I'm not responsible for the design part. All I can say is that B *also* serves as a backup MX. There's also C somewhere for real redundancy. Yes, we got our users in LDAP. DSPAM uses MySQL. Actually that was the main factor we decided we need to use the same DSPAM instance for B, because there's mail for server A users, and their filters are already trained. -- () Campania Panglicii în ASCII /\ http://www.asciiribbon.org/ ------------------------------------------------------------------------------ _______________________________________________ Dspam-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/dspam-user
