Hello,
I have a mail server with postfix 2.5.5 (Debian lenny) which is serving mail
forwarding domains as described in VIRTUAL_README.html without any problems.
The respective part in main.cf contains (domain name removed):
virtual_alias_domains = xxxx
virtual_alias_maps = cdb:/etc/postfix/virtual
Now I had to set up a SASL-authenticated (just using sasldb) submission
instance on that server on a non-standard port (25025) for a couple of users.
The respective section in master.cf then reads like this:
smtp inet n - - - - smtpd
25025 inet n - - - - smtpd
-o smtpd_tls_auth_only=no
-o smtpd_sasl_auth_enable=yes
-o smtpd_sasl_security_options=noanonymous,noplaintext
-o smtpd_sasl_local_domain=xxxx
-o broken_sasl_auth_clients=yes
-o smtpd_client_restrictions=permit_sasl_authenticated,reject
-o
smtpd_recipient_restrictions=permit_sasl_authenticated,reject_unauth_destination
-o content_filter=
With this the standard mail server still works like before and ESMTPA relaying
for the authenticated users also works like expected but for one case: it
bounces for recipients with addresses in the virtual forwarding domains which
should be handled (i.e. mapped) by the server itself.
I have tried to add the virtual_alias_domains and virtual_alias_maps settings
explicitly via "-o" overwrites to the 25025 service in master.cf, but to no
avail. Turning on verbose messages with "-v" on the 25025 smtpd shows in the
logs that this smtpd is able to correctly resolve the virtual names in the
section
>>> CHECKING RECIPIENT MAPS <<<
...
postfix/smtpd[11296]: maps_find: virtual_alias_maps:
cdb:/etc/postfix/virtual(0,lock|fold_fix): test@xxxx = test@yyyy
postfix/smtpd[11296]: mail_addr_find: test@xxxx -> test@yyyy
but mail to the address test@xxxx still bounces if submitted via the smtpd
on port 25025 while at the same time it is successfully delivered if injected
via the standard smtpd.
I assumed that for the 25025 smtpd anything which is not overwritten with "-o"
in master.cf is exactly the same as defined in main.cf, but obviously I must
be missing something here. Any hint as to what it is is highly appreciated.
Kind regards,
Ch. Scheurer
