Postfix 2.3.3 on CentOS 5.8
: In a dual-stack server with Postfix using both IPv4 and IPv6, sending to a certain IPv4 only domain - with no AAAA DNS record - fails
: Postfix tries IPv6 first and then fails, with no IPv4 attempt
: status=bounced (Host or domain name not found. Name service error for name=DOMAIN.com type=AAAA: Host not found)
I have sanitized the log entries below, but if some information is necessary in determining what is happening, I can communicate the relevant domain directly to you.
For what it is worth, main.cf is set to use both IPv4 and IPv6, and the server has native IPv4 and IPv6 internet connections:
inet_protocols = all
And this is the log snippet - first with inet_protocol = all, then changed to inet_protocol = ipv4 and postfix restarted:
Feb 2 18:36:33 Calevra postfix/smtp: 05C6D285802C: to=<user@DOMAIN.com>, relay=none, delay=5.8, delays=0.46/0.02/5.3/0, dsn=5.4.4, status=bounced (Host or domain name not found. Name service error for name=DOMAIN.com type=AAAA: Host not found)
Feb 2 18:45:57 Calevra postfix/postfix-script: stopping the Postfix mail system
Feb 2 18:45:57 Calevra postfix/master: terminating on signal 15
Feb 2 18:45:57 Calevra postfix/postfix-script: starting the Postfix mail system
Feb 2 18:45:57 Calevra postfix/master: daemon started -- version 2.3.3, configuration /etc/postfix
Feb 2 18:46:49 Calevra postfix/smtp: 50D6A285802B: to=<user@DOMAIN.com>, relay=smtp.secureserver.net[188.8.131.52]:25, delay=1.6, delays=0.45/0/0.59/0.57, dsn=2.0.0, status=sent (250 ok: Message 1004841767 accepted)
Anyone have any ideas why postfix is starting with IPv6, and doesn't fail back to IPv4 if there is a send failure? Is there some sort of domain misconfiguration that could cause this, a Postfix problem, or a GoDaddy mail server (smtp.secureserver.net) problem?!