Thursday, March 14, 2013

After changing the SMTP connector, Error: 451 4.4.0 Primary Target IP address responded 421 4.2.1 unable to connect.

In the last week I was involved in the following issue:
After changing the configuration of the SMTP send connector on a NLB cluster (second smart host added) user reported that no mails can be sent to the internet. The other way round (receiving) was working fine.


Configuartion:

Exchange Server 2007 - Hub Transport role, Send Connector with the following configuration:
All SMTP outgoing mails are sent to a smart host in the perimeter network.

Error Message:

Checking the send-queue on the exchange server, we see there over 70 emails that are not sent with the following error message:

451 4.4.0 Primary Target IP address responded 421 4.2.1 unable to connect.  
Attempted failover to alternate host, but that did not succeed. 
Either there are no alternate hosts or delivery failed to the all alternate host.
Troubleshooting steps:
  • We reviewed the smarthost server and logs but we didn’t  find any errors.
  • Incoming mails were forwarded without any problems to the exchange server.
  • Attempting to send a test mail using the telnet port 25 directly to the smarthost went well.
  • We have set the logging to verbose mode on the exchange server and smarthost.
  • After checking the log file on the smarthost server we saw no attempt for a connection from the exchange server to the smarthost. Just our e-mail sent by telnet could be seen in the logfile.
  • In the log file on the exchange server merely “451 4.4.0 Primary Target IP address responded 421 4.2.1 unable to connect....”   is shown.

  
Solution:
We decided to create new send connector for testing with just one special test domain (gmx.net) and a cost with a value as high as the other  connector ( =1 ).
After sending a  test-mail to our test recipient (mausi@gmx.net) were successful.
E-mail was transferred without any problem to our smart host and sent to the recipient.

Conclusion:  send connector object has been corrupted!

We put in a not existing domain to our old connector and set costs to a maximum value of 99 to deactivate it to be deleted after some days of monitoring. Our newly created connector got the original value for the domain (*) and the cost ( 1 ).

Sequential problem:
In the meantime we had a lot of e-mails in the queue still waiting for transmission.
Now the important question: How to get all the e-mails from the old queue to the new one?

No problem – restarting the transport service brought them to the correct queue and all mails were delivered immediately to the smarthost server and sent out.

Recommendation: 
create second send connector with high costs for alternate/backup way.





3 comments:

  1. Did the issue ever come back?
    Cause i have the exact same issue and when restarting the transport service (without creating a new send connector) the mails go out too. The problem returns after a couple of hours though.

    I've created a new send connector as suggested by you just a few seconds ago. Im wondering if this solves this problem permantly.

    Can you let me know if the issue returned at your side?

    ReplyDelete
    Replies
    1. sorry for late reply the issue returned not more again.

      Delete