> f it is "always_bcc" you have for now,
> then you should indeed enable it only on the far side of the content filter,
> which amounts to:
>
> -o receive_override_options=no_address_mappings
This was one of the things I already tried without resolving the issue, as
stated in my feedback mail:
> 6) Added the following entry to the master.cf
> smtp inet n - - - - smtpd
> -o content_filter=spamassassin
> ->> -o receive_override_options=no_address_mappings
________________________________________
Von: owner-postfix-us...@postfix.org [owner-postfix-us...@postfix.org]" im
Auftrag von "Viktor Dukhovni [postfix-us...@dukhovni.org]
Gesendet: Montag, 5. Februar 2018 21:59
An: Postfix users
Betreff: [MASSMAIL]Re: Duplicate mails in mailq / always_bcc
> On Feb 5, 2018, at 3:16 PM, Viktor Dukhovni <postfix-us...@dukhovni.org>
> wrote:
>
> One thing that is interesting is that the message generates two BCC
> copies with the *same* recipient address:
>
> Feb 5 17:28:51 mail postfix/smtp[22771]: 363BD601EE:
> to=<maildepot@mailappliance.local>,
> relay=192.168.1.23[192.168.1.23]:1025, delay=0.1,
> delays=0.01/0/0.04/0.04, dsn=4.4.2,
> status=deferred (lost connection with 192.168.1.23[192.168.1.23]
> while sending end of data -- message may be sent more than once)
>
> Feb 5 17:28:51 mail postfix/smtp[22771]: 363BD601EE:
> to=<maildepot@mailappliance.local>,
> relay=192.168.1.23[192.168.1.23]:1025, delay=0.1, delays=0.01/0/0.04/0.04,
> dsn=4.4.2,
> status=deferred (lost connection with 192.168.1.23[192.168.1.23]
> while sending end of data -- message may be sent more than once)
This is very likely a result of having "always_bcc" enabled on both sides of
the content filter. So while "recipient_bcc_maps" is generally a better way
to archive email than "always_bcc", if it is "always_bcc" you have for now,
then you should indeed enable it only on the far side of the content filter,
which amounts to:
-o receive_override_options=no_address_mappings
in master.cf for the smtpd(8) service that accepts the original message
for handing off to amavis. This should generate new BCC messages with
just one instance of the BCC recipient. Post log observations for those,
and any relevant tcpdump data (for just one extracted TCP session that
fails, not everything that comes off the wire at the time).
--
Viktor.
