Postfix 3.1 problems

Hello.

I’ve had RT working for quite a while on Ubuntu 14.04 with Postfix 2.11.0.

I’ve since spun up a new Ubuntu instance (16.04) which comes with Postfix 3.1. RT installed just fine and most things seem to be working. Incoming email works fine on my older install (i.e., postfix picks it up and sends it to the proper queue) – but this does not work on my new install with Postfix 3.1.

As a note – I simply am not a Postfix master, I barely know how to spell it! At this point everything looks the same between my two installations from a configuration perspective. I’m left to thinking that Postfix 3.x is my problem. Looking at /var/log/mail.log I see:

Aug 7 14:06:36 rt-hr postfix[10811]: Postfix is running with backwards-compatible default settings
Aug 7 14:06:36 rt-hr postfix[10811]: See http://www.postfix.org/COMPATIBILITY_README.html for details
Aug 7 14:06:36 rt-hr postfix[10811]: To disable backwards compatibility use “postconf compatibility_level=2” and “postfix reload”

So I did those two postfix commands on a whim and it still does not work.

When sending test emails to my RT queues I do not get a bounce back for many many hours; but eventually I get:

The response was:
The recipient server did not accept our requests to connect.

The system is resolvable from the Internet – I do not have an MX record setup though – I have an A record.

With that little information, does anybody have any suggestions?

Thanks!

If you’re trying to send mail to rt at the domain instead of at the hostname (say, rt@mydomain.com instead of rt@rtserver.mydomain.com) you’ll need to have the MX record set up.

postfix might be set up in send-only mode; check that it’s configured to receive mail as well, and that it’s not set to listen only on the loopback interface (see https://www.digitalocean.com/community/questions/postfix-mail-forwarding-socket-error-how-to-fix-mx-records). Check that iptables is allowing incoming mail. If you can send to root or postmaster at your rt server from a different system then these things should be okay.

Find the postfix log (probably in /var/log/postfix) and find out what it says it did with the message, if it saw it at all. Find the rt log (maybe /var/log/rt.log) and find out if it saw the message.

Good luck,
Kyle

Thanks for the reply!

I am trying to send email to a setup such as rt@rtserver.mydomain.com – so I think an A record should suffice…correct?

Also I have the inet interfaces = all in my main.cf file.

UFW is currently inactive. I can see postfix is listening on port 25:

root@host:/etc/postfix# netstat -plutn | grep 25
tcp 0 0 0.0.0.0:25 0.0.0.0:* LISTEN 10848/master
tcp6 0 0 :::25 :::* LISTEN 10848/master

There is nothing pertaining to email in the rt.log file; I can’t seem to hunt down the specific postfix log – that’s why I was looking at /var/log/mail.err and mail.log…

I’m still scratching my head here… :slight_smile:

All of my google searches today say that postfix logs to /var/log/mail.log. Sorry for the red herring (I use exim, not postfix). Check the “mail.*” line in /etc/rsyslog.d/50-default.conf to be sure, though.

Does the log have any information about incoming messages at all?

Maybe some of the steps here:

http://www.postfix.org/DEBUG_README.html#debug_peer

can help.

You can also try to send a message with telnet:

telnet rtserver.mydomain.com 25
HELO
MAIL FROM:
RCPT TO: rt@rtserver.mydomain.com
DATA

test message

.

then ctrl-] to get back to the telnet prompt, and quit.

If it’s working you’ll have a conversation that ends with a 250 code after the single ‘.’ on the last line. If it’s not working you’ll have more information to work with, depending on where it stopped working. If it never responds to telnet at all, a firewall somewhere is blocking port 25.

Try the telnet thing both from the rtserver, and from somewhere else.

Sending mail to rt@rtserver.mydomain.com should be fine with an A record, correct.