this time I'm not stuggeling with a client, but with my own setup. We are running about 10 mailboxes off a 2013 Exchange server. It has an external name and a single domain certificate. Access is configured using split-brain DNS-configuration. External autodiscover uses the SRV method. All is well.
Even though I have migrated several Exchange 2010 to 2016, I have never migrated one from 2013.
I set up the new box, added it to the domain, installed Exchange 2016 and gave it a different external name, added that to the DNS also in split-brain configuration.
I migrated several boxes, all was well. No changes were made to the configration of the old server. OWA works on both servers just fine.
Outlook cannot access either machine anymore. That's not completely right - it works off our RDP server using Outlook 2013 with exchange proxy values manually added. I am getting several certificate errors on the local physical machine name, even though this has never ben the case in the past. Outlook freezes if it connects at all. New profiles don't find their way to the 2016 server using Outlook 2016 - the message says it could not log on to the server because the database is down (which it is not). Smart phones work fine.
My question is this:
Where do I point the autodiscover SRV-record? To the first or to the second server or does that not matter?
Where do I point the autodiscoverinternaluri-value? Each server to itself? That's the way I did it right now.
I can't find anything wrong and the logs are fine. What can I do? I hesitate to just migrate the remaining boxes and uninstall 2013 - I want to understand what is happening...