On Tue, Apr 19, 2011 at 11:06:07AM -0500, Jay G. Scott wrote:
> i just swapped postfix in replacing a sendmail configuration.
>
> here is how i thought it worked (but i'm wrong):
>
> my relay recipient map has entries like this:
> g...@arlut.utexas.edu OK
>
> my transport file has entries like this:
> ns8.arlut.utexas.edu smtp:[ns8.arlut.utexas.edu]
>
> i thought this meant that an address of the form:
> g...@ns8.arlut.utexas.edu
> would get delivered, but it's bouncing.
> user unknown in relay recipient table
arlut.utexas.edu != ns8.arlut.utexas.edu
> i was hoping that, as long as transport_maps had an entry for
> every machine that could take delivery of mail that an address
> of the form:
> user@<some entry in transport maps>
> would be allowed.
> i guess that's not the case? is there something i can do that
> will achieve what i want w/o the massive duplication in the
> relay recipient table?
Sounds like what you might need is virtual_alias_maps to rewrite
g...@arlut.utexas.edu to g...@ns8.arlut.utexas.edu. Then your transport
mapping would get that mail to ns8.arlut.utexas.edu. (Or better yet
just leave off the transport_maps, because apparently you have DNS
for ns8.arlut.utexas.edu.) (I show MX records for ns8, so you might
want to change those to point to where you want the mail sent.)
> i can (and have) generated a list of
> all the real users, and i have them in the relay recipient map.
> does that map have to have an entry for every machine that a
> given user might get mail on? because that's going to be much
> harder for me to accomplish.
No, the relay_recipient_maps are a list of all valid addresses in
relay_domains. See
http://www.postfix.org/ADDRESS_CLASS_README.html#relay_domain_class
(second time today I have posted that link :) )
> i'm adding my postconf output, but i think my problem is more
> fundamental.
Yikes! The list guidelines ask for "postconf -n" *not* full postconf
output! I didn't read that.
--
Offlist mail to this address is discarded unless
"/dev/rob0" or "not-spam" is in Subject: header
