none
External e-mail issue following Exchange 2013 CU12 upgrade

    Question

  • Hi everyone

    We recently upgraded our two Exchange 2013 servers to CU12, but since then we have had some issues with incoming external mail being delayed.  Our Exchange setup is below:

    MailServer1 - CAS & mailbox

    MailServer2 - mailbox

    Our public DNS points to our CAS server through a firewall so all external mail is routed to the CAS server.  Before reaching the firewall, mail goes through a McAfee smart host.  Since the upgrade to CU12 we've noticed a large number of e-mails stopping at the smart host with the error:

    4.7.0 Temporary server error. Please try again later. PRX4

    The e-mails keep trying to send and eventually get through, but this is causing some external e-mails to be delayed by up to 2 hours.

    I have checked the submission logs (Exchange Server\V15\TransportRoles\Logs\Mailbox\Connectivity\Submission) on the CAS server and this error also appears a number of times here.  Another strange thing is that when mail is accepted from the smart host it only shows MailServer2 as the responding server on the McAfee portal, whereas before the CU12 upgrade either of the 2 could respond.

    I've checked the event logs on both mail servers and the only errors shown on there relate to performance counters not being loaded, which I don't think is related.  When I check the queues on the two mail servers, only a submission queue is shown on MailServer1, where MailServer2 shows a "SmartHostConnectorDelivery" queue and "SMTP Delivery to Mailbox" queues.  I'm sure that prior to the CU12 upgrade these extra queues were on MailServer1 rather than 2.

    I'm not quite sure where to go from here as mail is working generally but the external email delays are becoming a problem.  I'd be grateful if someone could let me know of anything that may have caused this problem, and whether anyone else has experienced similar issues after upgrading to CU12?

    Thanks for your help in advance.


    • Edited by gjayne Wednesday, April 20, 2016 10:14 AM
    Wednesday, April 20, 2016 9:37 AM

All replies

  • Hi everyone

    We recently upgraded our two Exchange 2013 servers to CU12, but since then we have had some issues with incoming external mail being delayed.  Our Exchange setup is below:

    MailServer1 - CAS & mailbox

    MailServer2 - mailbox

    Our public DNS points to our CAS server through a firewall so all external mail is routed to the CAS server.  Before reaching the firewall, mail goes through a McAfee smart host.  Since the upgrade to CU12 we've noticed a large number of e-mails stopping at the smart host with the error:

    4.7.0 Temporary server error. Please try again later. PRX4

    The e-mails keep trying to send and eventually get through, but this is causing some external e-mails to be delayed by up to 2 hours.

    I have checked the submission logs (Exchange Server\V15\TransportRoles\Logs\Mailbox\Connectivity\Submission) on the CAS server and this error also appears a number of times here.  Another strange thing is that when mail is accepted from the smart host it only shows MailServer2 as the responding server on the McAfee portal, whereas before the CU12 upgrade either of the 2 could respond.

    I've checked the event logs on both mail servers and the only errors shown on there relate to performance counters not being loaded, which I don't think is related.  When I check the queues on the two mail servers, only a submission queue is shown on MailServer1, where MailServer2 shows a "SmartHostConnectorDelivery" queue and "SMTP Delivery to Mailbox" queues.  I'm sure that prior to the CU12 upgrade these extra queues were on MailServer1 rather than 2.

    I'm not quite sure where to go from here as mail is working generally but the external email delays are becoming a problem.  I'd be grateful if someone could let me know of anything that may have caused this problem, and whether anyone else has experienced similar issues after upgrading to CU12?

    Thanks for your help in advance.


    What does Get-ServerComponentState show for each server? Ensure everything is set to "active"

    Also verify available disk space on the Exchange Servers with the transport queue is.


    Blog:    Twitter:   


    Wednesday, April 20, 2016 11:47 AM
  • Hi Andy, thanks for the quick reply.

    Get-ServerComponentState shows everything active on MailServer1.  On MailServer2, everything is active apart from ForwardSyncDaemon and ProvisioningRps. Do these need to be activated on MailServer2 also?

    Free space on both servers is fine with 85GB free on MailServer1 and 245GB free on MailServer2.

    Wednesday, April 20, 2016 11:53 AM
  • Hi Andy, thanks for the quick reply.

    Get-ServerComponentState shows everything active on MailServer1.  On MailServer2, everything is active apart from ForwardSyncDaemon and ProvisioningRps. Do these need to be activated on MailServer2 also?

    Free space on both servers is fine with 85GB free on MailServer1 and 245GB free on MailServer2.

    They do, but that wouldn't cause this issue.

    What I would do for now is remove the problematic server from the external mail flow so inbound only uses the good one. Is internal mail flow affected?

    Anything in the event logs? What you are seeing is typically resource related on the affected server. I assume its been bounced and all services are running. any 3rd party products installed? anti-malware etc...

    All connectors enabled? Can you post what you have set for the receive connector used to receive mail from the smarthost?


    Blog:    Twitter:   

    Wednesday, April 20, 2016 12:06 PM
  • I've checked the SMTPReceive logs on the server and it seems that there are some internal messages also being affected, however it seems the majority are external e-mails.

    We have 5 enabled receive connectors, all receive mail from any remote address and bind all NICs to the ports shown below:

    • Client Frontend - 587
    • Client Proxy - 465
    • Default Frontend - 25
    • Default - 2525
    • Outbound Proxy - 717

    We also have two custom connectors, which accept mail from a small number of IPs on port 25.  I read something about connectors that conflict with other connectors using the same ports causing issues but these two custom connectors are currently disabled.  I assume that as they're disabled they wouldn't cause any issues but do you think it's worth deleting them completely?

    Wednesday, April 20, 2016 2:20 PM
  • I also forgot to say we do have AV on the servers - Microsoft Forefront Protection.  However this is on both mail servers and causes no issues on MailServer2.  The AV is configured to exclude all the Exchange files etc when doing scans.

    I also tried disabling the AV on MailServer1 to test but the issues remains.

    Wednesday, April 20, 2016 2:27 PM
  • After a great deal of comparison between the two servers I've only been able to find one difference, which raises more questions than answers.  On the server that is doesn't seem to be functioning properly the tickbox for IPv6 is ticked as shown below:

    The server that is working ok doesn't have this tickbox ticked.  I find this confusing because everything I've read says that disabling IPv6 should be done via the registry, if at all, and can cause issues with Exchange 2013.  Our 2 exchange servers are on ESXi hosts and the hosts themselves also have IPv6 disabled so I'm not sure if the fact that IPv6 is enabled on the faulty Exchange server but not on the hosts is causing any issues?

    I was going to try disabling IPv6 on the faulty Exchange server to see if this resolves the issue but wanted to check whether this could cause further issues?

    Monday, May 16, 2016 3:38 PM
  • So, probleme solved? If yes how?
    Monday, July 3, 2017 8:31 AM