Re: question about receive_override_options

2 msg	postfix + gmail + multiusers
2 msg	strange question regarding lmtp
3 msg	Updated amavisd-new; can't send/receive mail
8 msg	RPM 'n stuff
7 msg	DSN setup
4 msg	Local Users on Virtual Domains?
5 msg	postfix + courier + active directory 2003
12 msg	Fail to test postfix
2 msg	Global groups
7 msg	About creating database
3 msg	Misunderstanding of transport map format VS mas...
2 msg	Problems Backscatter

question about receive_override_options
\ Peter Rabbitson (14 Jun 2007)
. \ MrC (14 Jun 2007)
. . \ Peter Rabbitson (14 Jun 2007)
. . . \ MrC (14 Jun 2007)
. \ mouss (15 Jun 2007)

4 msg	specific whitelisting
3 msg	Real unknown host ???
2 msg	Another question about logs...
4 msg	fallback_relay not getting used
3 msg	Maildir
6 msg	Outgoing Mail Copy
1 msg	Close to getting SASL and SMTP AUTH on OS X wor...

Subject:	Re: question about receive_override_options
Group:	Postfix-users
From:	mouss
Date:	15 Jun 2007

Peter Rabbitson wrote:
> Hello,
>
> I use an external scanner ( content_filter = scanner:[localhost]:10025
> ) and I am having a problem with duplicate messages triggered via
> sender_bcc_maps. The fine manual explains why the double-expansion is
> occurring, and that I need to set
> receive_override_options=no_address_mappings before I send mail to the
> content filter. What I am confused about is how exactly should I do
> this. Do I:
>
> A: simply set receive_override_options=no_address_mappings in
> main.cf, so all smtpd/pickup daemons will inherit it automatically,
> and only my scanner back-feed smtpd will perform the mapping as it
> already has -o
> receive_override_options=no_header_body_checks,no_unknown_recipient_checks,no_mi lters
>

that's ok, but you'll need to keep that in mind if you ever add an smtpd
listener for a special purpose.
>
> or
>
> B: I walk through master.cf and stick -o
> receive_override_options=no_address_mappings on every smtpd/pickup
> service

That's less "error prone" (see above).

if you have too many listeners and you feel "confident", consider using
multiple instances instead.