Exchange 2013 POP3 service drops connections fix


At my current client they are using POP3 and we are in the process of setting up Exchange 2013 servers to act as hybrid servers for their Office 365 migration.

So after setting up the servers I tested all protocols, then a day or so later I noticed that POP3 was showing down on the Kemp NLB. When I did TELNET test to port 110, from a remote machine, using the FQDN or IP the connection was dropped after a few moments, without returning any text.

I then tried from the Exchange server [2013 CU5, multi-role, 2 NICs (iSCSI and Pubic), POP bindings 0.0.0.0] itself and had the same results. I then tried 127.0.0.1 and localhost with TELNET and those worked. So the service was working, but not as expected.

I tried rebooting and resetting several settings on the PopSettings and searching for this issue on-line, but came up empty. So finally went to the TechNet Exchange Server 2013 – Outlook, OWA, POP, and IMAP Clients forum and searched on “pop connection” and found this post: POP works via localhost but not from other networked machines, which included the “fix.”

The issue was that the “State” of the PopProxy component was set to Inactive:

[PS] D:\>Get-ServerComponentstate -Identity SRVDENEX01

Server Component State
—— ——— —–
SRVDENEX01.company.com ServerWideOffline Active
SRVDENEX01.company.com HubTransport Active
SRVDENEX01.company.com FrontendTransport Active
SRVDENEX01.company.com Monitoring Active
SRVDENEX01.company.com RecoveryActionsEnabled Active
SRVDENEX01.company.com AutoDiscoverProxy Active
SRVDENEX01.company.com ActiveSyncProxy Active
SRVDENEX01.company.com EcpProxy Active
SRVDENEX01.company.com EwsProxy Active
SRVDENEX01.company.com ImapProxy Active
SRVDENEX01.company.com OabProxy Active
SRVDENEX01.company.com OwaProxy Active
SRVDENEX01.company.com PopProxy Inactive
SRVDENEX01.company.com PushNotificationsProxy Active
SRVDENEX01.company.com RpsProxy Active
SRVDENEX01.company.com RwsProxy Active
SRVDENEX01.company.com RpcProxy Active
SRVDENEX01.company.com UMCallRouter Active
SRVDENEX01.company.com XropProxy Active
SRVDENEX01.company.com HttpProxyAvailabilityGroup Active
SRVDENEX01.company.com ForwardSyncDaemon Active
SRVDENEX01.company.com ProvisioningRps Active
SRVDENEX01.company.com MapiProxy Active
SRVDENEX01.company.com EdgeTransport Active
SRVDENEX01.company.com HighAvailability Active
SRVDENEX01.company.com SharedCache Active

A quick call to Set-ServerComponentState to mark this component Active fixed the issue:

Set-ServerComponentState -Identity SRVDENEX01 -Component PopProxy -Requester HealthAPI -State Active

After doing this POP started responding as expected, using any valid hostname or IP address.

Like the user who posted to TechNet, I would also like to know why the health check failed and marked the PopProxy as inactive. I will research this further and if I find anything I will update this post.

About jasonsherry

I am a 20 year Exchange consultant and expert. I currently work for Commvault as a Solutions Specialist for Microsoft Infrastructure For more info see my resume at: http://jasonsherry.org
This entry was posted in Exchange, Microsoft, Technical and tagged . Bookmark the permalink.

37 Responses to Exchange 2013 POP3 service drops connections fix

  1. Chris Zimmerman says:

    I have been having the EXACT SAME issue! I have 2 servers in a cluster, and one of the two (always the same one) dies by going inactive by the HealthAPI. I have to force it back up all the time. No clear reason why. If you find anything, I would appreciate the help, as i’ve been tearing my hair out trying to figure out this one.

    Like

  2. Jesper Sandberg says:

    Thank you a lot for sharing this info. Saved me hours debugging

    Like

  3. Pingback: Fixed POP3 Service after Exchange 2013 CU6 broke it | jermsmit.com

  4. Jermal Smith says:

    Hey Thank you so much Jason, you saved me some time in waiting for Microsoft product support.
    I have just finished dealing with them over a replication issue caused by CU6. This issue happened to us right after our CU6 upgrade which we were instructed to due to address a Hybrid issue

    I gave you credit for this in my blog post: http://jermsmit.com/fixed-pop3-service-after-exchange-2013-cu6/

    Once again, thank you

    Like

  5. Have you found the cause of its failure yet?
    We have this popproxy failing each day at random hours, not able to find a cause?

    Like

    • jasonsherry says:

      No, this was for a client that used POP3. Once we fixed the setting with Set-ServerComponentState the issue didn’t come back up.

      Open a ticket with support\PSS and they will track it down and it won’t cost you anything; your money or issue use will be refunded.

      Like

  6. Paul says:

    Thank you. Very helpfull info. I had the same issue.

    Like

  7. Mumen says:

    Thank you very much for this valuable information

    Like

  8. Scott Cuff says:

    This solved the login to pop but now the smtp fails. Oddly I have a vpn and use the same settings 587 tls and it works from my desktop from from anywhere outside the send test fails. I definitely have port 587 open on the firewall.

    Like

    • jasonsherry says:

      This post is about POP3, 587 is (normally) used by SMTP. In any case, a blog isn’t a good support form. If you post your issue on the TechNet forums you have a much better chance to get help there.

      Like

  9. Eric Acosta says:

    Thank you very much! This fixed my issue with the ImapProxy component. Thanks again!

    Like

  10. Erik Nettekoven says:

    You saved my day with this one! We are running a clean installation of Exchange 2013 CU6, the problem occurred after we installed KB3011140 for Exchange 2013 CU6. As soon as we ran the command from your post, POP started to work again.

    Another error that might point to the problem described is if you encounter this (multiple times) in the pop3 logging:
    127.0.0.1:995,127.0.0.1:37195,,1,0,31,CloseSession,, (maybe add it to your blogpost?)

    Like

  11. jasonsherry says:

    I don’t see an error in that log entry. Just looks like connection info, unless the last “1,0,31” part is an error code.

    Like

  12. Roland Lyngvig says:

    I ran into exactly this problem. I have nearly pulled out all hairs on top of my head to solve this problem. The problem started here after I upgraded the Exchange certificate. I may have deleted the old one in the wrong order or so. But Jason, you’ve saved my day. POP3 and IMAP (which had the same problem) are now working again.

    Like

  13. Tobie Fysh says:

    Thanks Jason, was causing us problems and your post helped greatly.

    Like

  14. erpcoder says:

    Thank you for this post, this solved my problem

    Like

  15. Brian says:

    So all of these components should be active? I still have 3 more that are inactive. BTW this corrected my issue, thanks!

    Like

  16. Pavel says:

    Thank you a lot for sharing this very useful info! I lost my weekend ….

    Like

  17. Boyd says:

    Thanks Jason after 3.5 days trying to fix my POP3 connection I found this post and now it is all ok.
    You saved me just in time for a customer demo tomorrow 🙂

    Like

  18. Anderson Zardo says:

    We have CU7 here and the issue stil occurs. I’m sitting here waiting Microsoft to fix that….

    Like

    • Brian says:

      In our case, editing the server settings corrected the component issue. Make sure that POP3 and IMAP4 have (All available IPv6) and IPv4 listed instead of the servers actual IP. This goes for both the TLS and SSL connection settings. Once I put this back in place the components stopped going offline.

      Like

      • Anderson Zardo says:

        Hi Brian, ipv4 settings are configured for the server’s actual IP for 995. Changed it to “all available ipv4”. Let’s see…
        Thanks for the tip.

        Like

  19. Marc says:

    Fixes out problem after a netwerk broadcast storm.

    Like

  20. Thang Tran Duc says:

    I did run commands as Jasons Herry suggested but POP3 and POP3backend services failed to start. The commands that I run as follows:
    Get-ServerComponentstate -Identity Servername
    Set-ServerComponentState -Identity Servername -Component PopProxy -Requester HealthAPI -State Active
    Any idea?

    Like

  21. markfe says:

    we are using pop3 for some voicemail boxes of a broadsoft voip solution, and someday pop3 just stopped working… because of 0 logging but for this error:
    after: The service has started successfully. PID 12100 EVENTID 1001
    there came this: The POP3 server’s Online status has changed to False. EVENTID 1011

    i looked hours for a solution ,because eventlog was full of errors regarding perfcounters i looked there first but could find anything: thanks a lot for this !

    i guess this came from too many connections on port 110 …

    Like

  22. Manjula says:

    Hi,
    Thanks a lot i was searching the related issues, all over..

    You saved the day of mine.. anyway i copy the error msg i got from pop clients. it will help full for other who doing google.

    Unable to send or receive messages for the xx account. Your server has unexpectedly terminated the connection. Possible causes for this include server problems, network problems, or a long period of inactivity. Server: ‘POPServer’ Windows Live Mail Error ID: 0x800CCC0F Protocol: POP3 Port: 995 Secure(SSL): Yes

    Like

  23. magviegas says:

    It seems that the Hybrid Configuration with Office 365 is the responsible for that. In my case it was.

    Like

  24. Robert Richter says:

    You saved my life 🙂 ! BIG THANK YOU!

    Like

  25. Felipe das Chagas says:

    My name is Felipe from Brazil. Thank you for the post. Same problem fixed in my company.

    Like

  26. Alex Se says:

    THANK YOU!!!!

    Like

  27. Dariusz Sroka says:

    THANKS!

    Like

  28. Insheina says:

    Thank you! I had a same problem after installing Exchange 2013 CU13.

    Like

  29. Mhamed says:

    From Morocco.
    Great post; thanks !!!!

    Like

  30. Alex says:

    Thanks very much. Scratching my head after an hour or so and found this post. Was pretty much out of ideas after restarting the POP services on both CAS and MBX servers.

    Like

  31. Thomas says:

    Thanks. exact same problem, you solved it with your post. thanks Thomas

    Like

  32. You Ki Seung says:

    Your knowledge has been a great help to me.

    Like

  33. Dan Chandler says:

    Thank you so much for this! I spent hours trying to figure out what was causing the issue with one of our internal applications failing to send out emails – this worked a treat!

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s