On 11/14/2017 6:00 AM, flowhosts wrote: > Hello, > > as described in the subject i tried to implement the new feature > check_recipient_a_access > I have encountered a strange error or maybe an bug. > > The following settings result in an correct action follwed by an > "4.3.5 Server configuration error" response. > # main.cf > smtpd_recipient_restrictions = > reject_non_fqdn_sender > ... > check_recipient_a_access > hash:/etc/postfix/lookup/recipient_a_access > ... > permit > > # cat /etc/postfix/lookup/recipient_a_access > 185.140.110.3 DISCARD > > # maillog > Nov 14 10:53:54 fallback postfix/smtpd[7187]: NOQUEUE: discard: RCPT > from unknown[192.168.xxx.xxx]:53698: > <vasilnhdgz0sdiminut...@netgooya.com>: Recipient address triggers > DISCARD action; from=<> to=<vasilnhdgz0sdiminut...@netgooya.com> > proto=ESMTP helo=<bsmtp.xxx.xx> > Nov 14 10:53:54 fallback postfix/smtpd[7187]: warning: restriction > check_recipient_a_access returns OK for > vasilnhdgz0sdiminut...@netgooya.com > Nov 14 10:53:54 fallback postfix/smtpd[7187]: warning: this is not > allowed for security reasons > Nov 14 10:53:54 fallback postfix/smtpd[7187]: warning: use DUNNO > instead of OK if you want to make an exception > Nov 14 10:53:54 fallback postfix/smtpd[7187]: NOQUEUE: reject: RCPT > from unknown[192.168.xxx.xxx]:53698: 451 4.3.5 Server configuration > error; from=<> to=<vasilnhdgz0sdiminut...@netgooya.com> proto=ESMTP > helo=<bsmtp.xxx.xx> > Nov 14 10:53:54 fallback postfix/cleanup[7844]: 3ybjWk29Jhz5vXS: > message-id=<3ybjwk29jhz5...@smtp.xxx.xx> > > > If DISCARD is replaced by HOLD in "recipient_a_access" the error > won't appear but in fact the sending host also receives an OK > message like it does above when discarding the mail, which should > not be allowed if you trust the warning message received. > > So is this a bug when using DISCARD or is it the right behaviour? > And if it's not a bug then i think HOLD is buggy because it does not > respond with an "451 4.3.5 Server configuration error". > > Where can i file a bug report? > Or can someone confirm this behaviour? > > Thanks in advance, > Patrick >
Confirmed (on postfix 3.2-20160730, using an inline: map). This looks like a bug, consider it reported. Thanks for finding this. -- Noel Jones