Rafael Azevedo:
> > Postfix always respects the MX preference order. If you believe
> > it does not, then YOU must show the tcpdump packet recording.
>
> I do believe that, but for some reason I'm getting this error when
> postfix for some reason seems to be sending to secondary MX servers.
> When I run postqueue -i MSGID it tries to delivery again and then
> mail gets sent.
Postfix logs all connection attempts in the maillog file including
al the ttempts that fail. A bit of logfile analysis will show that
it respects the MX order.
$ grep 'your error message' /var/log/maillog
That will show lines with a Postfix queue ID.
postfix/smtp[pid]: queueid: text including your error message
Then look for all logfile records with that pid:
$ grep postfix/smtp.pid /var/log/maillog
Example:
Jan 16 03:28:03 spike postfix/smtp[73083]: connect to
mwall.xxxxxxxx.com.au[xxx.29.91.92]:25: No route to host
Jan 16 03:28:19 spike postfix/smtp[73083]: connect to
mwall.xxxxxxxx.com.au[xxx.29.91.77]:25: No route to host
Jan 16 03:28:49 spike postfix/smtp[73083]: connect to
mwall.xxxxxxxx.com.au[xxx.29.91.93]:25: Operation timed out
Jan 16 03:34:02 spike postfix/smtp[73083]: 3YmM471q1Mzk2RR:
conversation with mwall2.xxxxxxxx.com.au[xxx.29.91.93] timed
out while receiving the initial server greeting
DO NOT turn on verbose logging. It will make the analysis much more
difficult.
Wietse
