I am having trouble in delivering mail to google from a fresh sendmail install on ubuntu 14.04. The DNS-records seem to be fine for the ip. Something else must be wrong.

Sending mail from command line:

sudo sendmail -v -Am -i [email protected];

Verbose output:

myname@fx1:/etc/mail$ sudo sendmail -v -Am -i [email protected];
[email protected]... Connecting to aspmx.l.google.com. via esmtp...
220 mx.google.com ESMTP v1si55415385wja.21 - gsmtp
>>> EHLO staging.mydomain.com
250-mx.google.com at your service, [2a01:4f8:212:27c8::2]
250-SIZE 35882577
220 2.0.0 Ready to start TLS
>>> EHLO staging.mydomain.com
250-mx.google.com at your service, [2a01:4f8:212:27c8::2]
250-SIZE 35882577
>>> MAIL From:<[email protected]>
250 2.1.0 OK v1si55415385wja.21 - gsmtp
>>> RCPT To:<[email protected]>
>>> DATA
250 2.1.5 OK v1si55415385wja.21 - gsmtp
354  Go ahead v1si55415385wja.21 - gsmtp
>>> .
550-5.7.1 [2a01:4f8:212:27c8::2] Our system has detected that this message does
550-5.7.1 not meet IPv6 sending guidelines regarding PTR records and
550-5.7.1 authentication. Please review
550-5.7.1  https://support.google.com/mail/?p=ipv6_authentication_error for more
550 5.7.1 information. v1si55415385wja.21 - gsmtp
myname... Connecting to local...
myname... Sent

Any help is greatly apreciated. Thank you in advance.

  • Have you reviewed the document in the error? Oct 28, 2015 at 10:31
  • yes of course. It states that the ipv6 dns must be set. I did so and the ipv6 adress points to the domain name.
    – merlin
    Oct 28, 2015 at 10:32
  • And your forward hostname is? Oct 28, 2015 at 10:35
  • not sure what you mean with forward hostname. The reverse DNS for 2a01:4f8:212:27c8::2 is set to staging.findix.com
    – merlin
    Oct 28, 2015 at 10:38
  • 4
    There's your problem. staging.findix.com does not have an AAA record pointing to 2a01:4f8:212:27c8::2. Oct 28, 2015 at 10:39

4 Answers 4


From Google's point of view, they are trying to verify the identity of the IP address connecting to them, so they will attempt to look up the PTR record for 2a01:4f8:212:27c8::2.

When they resolve that to staging.findix.com, they will then attempt to check that this resolves back to 2a01:4f8:212:27c8::2 - which it doesn't - see this lookup result.

Reverse (PTR record) Lookup

cwatson@thor:~$ nslookup

Non-authoritative answer:    name = tyr.vikingserv.net.

Forward (A record) Lookup

cwatson@thor:~$ nslookup tyr.vikingserv.net

Non-authoritative answer:
Name:   tyr.vikingserv.net
  • 6
    That first para isn't quite right; since multiple A/AAAA records can point to the same IP address, that would be impossible. What's required is that whatever hostname your PTR record points to, when you look that hostname up, you should get back to the original address. That is, the address->hostname->address loop must be complete; the hostname->address->hostname loop cannot be so guaranteed, and noone should care about it.
    – MadHatter
    Oct 28, 2015 at 10:54
  • Thank you. That resolved the problem and messages from the terminal get delivered.
    – merlin
    Oct 28, 2015 at 11:27
  • 3
    It seems to me you answer the "why" but not the "how to fix". I personally don't have enough DNS knowledge for this answer to help me.
    – Bob
    Apr 12, 2018 at 20:34
  • 2
    I had to add the AAAA record for my IPv6 address from the google error message (2a01:4f8:212:27c8::2 in your case) to the DNS server of the domain name returned by the hostname command. After some time (the expiry of the TTL which was the next day) I was able to send email to my gmal account with this command: sendmail -v -Am -i <my_gmal_account>@gmail.com; (and pressing Ctrl+D). Nov 2, 2018 at 13:50
  1. Check the current protocols:

    postconf inet_protocols

    net_protocols = all

  2. Edit cf File if it returns all

    grep 'inet_protocols' /etc/postfix/main.cf

    inet_protocols = ipv4

  3. Restart

    service postfix restart

  4. Check it again

    postconf inet_protocols

    inet_protocols = ipv4

  • 7
    IMHO a much easier solution than the accepted one. If you have no pressing need to use IPv6, this could be a quick fix. Jan 6, 2018 at 10:05
  • 1
    This is simpler than the other answer, although the other is more complete. It solved my problem as I don't need IPv6 for now.
    – javsmo
    Jul 15, 2019 at 19:52
  • 1
    This solved my problem. Why would I need IPV6? Jul 29, 2021 at 4:38

Your smtp server hostname could be a problem. It seemed to be the case for me. POstfix was somehow sending a host name to smtp.gmail.com not relevant with my smtp domain name.


I think Google's response is a bit disingenuous. Even though their "error message" mentions PTR, the 550 5.7.1 is about relaying. I too was getting this error on a Postfix / Dovecot install, but I fixed it via a few step:

  1. Adding the IPv6 IP info to your ifcfg-eth0 config on the mail server
  2. Adding a AAAA MX record on your DNS
  3. Adding the IPv6 ranges of all "authorized senders" in your postfix/main.cf on your mail server.

I have AT&T Business UVerse, and they don't support static IPv6 for PTR, but after doing the above steps Google stopped erroring out. I used https://www.ultratools.com/ipv6Tools to convert from IPv4 to IPv6. Finally, I put everything together in a howto on my site at Enabling IPV6 on CentOS (for sending and receiving mail)

  • What's an "AAAA MX record"? Surely you mean an AAAA record?
    – John
    Oct 9, 2022 at 15:11

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .