I have an Active Directory domain. We have two on-premises mail servers -- one running Exchange 2010 (exchange2010.example.com) and one running Exchange 2016 (exchange2016.example.com). My goal is to migrate from 2010 to 2016, but to have them in coexistence during the transition.

We have a firewall (pfSense) and allow access to OWA using NAT. I have one NAT rule pointing to the Exchange 2010 server and one pointing to Exchange 2016. From the outside, the IP for both servers look the same (since they are both running from the same internet connection), but there are two CNAME entries in external DNS, and the NAT is forwarding two different ports.

Internally, everything works. If I go to https://exchange2016.example.com/owa I get the 2016 OWA login page. I can log in. If the mailbox is on Exchange 2010, OWA proxies and I get 2010 OWA. Otherwise I get the 2016 interface.

Externally, I have the following urls: https://exchange2010.example.com/owa and https://exchange2016.example.com:8443/owa . This is where things get weird.

Say I visit https://exchange2016.example.com:8443/owa . I can see in the firewall logs that the client DOES connect to the Exchange 2016 server. But then there is an immediate 301 redirect to exchange2010.example.com, and I get the OWA login page for Exchange 2010, not Exchange 2016. This happens BEFORE I log in, so OWA does not know anything about where the mailboxes live. The domain name (exchange2016.example.com) stays the same, but I lose the port forward (8443) in the URL.

I can see that there is a redirect in the firewall and IIS logs, but I do not know what is causing it. I have the external URIs in my virtual directories set properly, I think.

Does anybody know what might be happening here? Where should I look for troubleshooting information? I do not feel the port forward is the problem, but I do not know how I would tell.

  • Have you considered directing external traffic to only the Exchange 2016 server?
    – joeqwerty
    Jan 24, 2020 at 3:33
  • Yes, I am looking into that. I cannot make the Exchange 2010 unavailable yet, though. Jan 24, 2020 at 4:02
  • If you direct external traffic only to the Exchange 2016 server and the user mailbox is on the Exchange 2010 server then the Exchange 2016 server should proxy the connection to the Exchange 2010 server just like it does for internal traffic, no? Why would it behave differently for external versus internal traffic?
    – joeqwerty
    Jan 24, 2020 at 12:41
  • In principle this should work, but we have had enough troubles with this migration that I want to try running both in parallel. Jan 28, 2020 at 22:31

1 Answer 1


How about ECP?

For the redirect issue, it’s recommended to refer to the workaround in below KB: Exchange Server 2013 or Exchange Server 2016 redirects to Exchange 2010 for OWA, Outlook on the Web, and ECP

In addition, here's a article to get in detailed about Client Connectivity in an Exchange 2010 and Exchange 2016 coexistence, for your reference: Client Connectivity in an Exchange 2016 Coexistence Environment with Exchange 2010

  • Yes, I looked at these. I do not think the first link resolves the issue. I think the second link is relevant in terms of proxying, but it does not address the issue of the 301 redirect. Jan 28, 2020 at 22:33

You must log in to answer this question.

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