Hello,
It looks that I'll have to implement SRS on one of machines I maintain.
The machine (debian) has multiple canonical domains, and some forwarding
done in virtual_alias_maps and some in users' .forward.
possibilities are perl-based srs, pysrs and postsrsd.
pysrs seems to support milter, which would be the easiest to integrate,
but it doesn't seem to be working with current version in debian.
postsrsd looks quite easy to install and configure (done that).
according to its docs in https://github.com/roehling/postsrsd#known-issues
the standard configuration overwrites all sender addresses (even not
forwarded), but there tricks described in:
https://github.com/roehling/postsrsd/discussions/76
indicate that it can be avoided by using some tricks.
Since we use relayhost I think I can avoid defining different cleanup
service and instead:
- route all outgoing e-mail from local addresses to relayhost using
sender_dependent_relayhost_maps (to it doesn't get SRSed)
- route all outgoing e-mail to alternative smtp port on localhost where
postsrs is used for sender_canonical_maps and mail is directed to
relayhost.
In addition I will
- incoming mail to srs0 BCCed to postmaster (so we quickly know about
forwarding problems)
- use recipient_canonical_maps for incoming mail to rewrite
now my questions:
0. Do you have better idea how to to it?
1. Do you see a potential problem here?
2. Do you have any ideas what to look for/avoid?
3. I'm curious, if I define TCP map that returns:
% echo 'get postmas...@example.com' | nc -N localhost 10001
500 Domain excluded py policy
- will this simply cause canonical maps not to hit? (500 looks like error)
--
Matus UHLAR - fantomas, uh...@fantomas.sk ; http://www.fantomas.sk/
Warning: I wish NOT to receive e-mail advertising to this address.
Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
I drive way too fast to worry about cholesterol.
