Arrgg! My Exchange 2013 migration was going so well until I added a new receive connector to enable a secure relay for our copiers etc..I added another IP to the internal NIC to handle this duty.
I made the change Wednesday night and didn’t notice anything wrong until Thursday morning when people started calling asking what’s up with the email. I started troubleshooting with my analyst while driving the 30 minute trip and tried the usual suspects. Check the queues, building up. Reboot, no help. Finally check the event viewer and notice the MSExchangeSubmission.exe is crashing repeatedly. Stumped at this point so I undid the change made the night before.
Check the queues again and notice they are failing on a DNS failure. Huh? I have no problems with my DNS servers, why can’t it resolve DNS queries. Command prompt: ping google.com…100% success rate. ping …0% success rate.
Check DNS, the server is listed and has the correct IP address. Getting frustrated now. Modified the hosts file and added the server/ip combo in it and everything is magically better. A couple hours later with everything running smooth, I decided to add my receive connector IP address back onto the NIC. Head into a meeting and start getting calls again. Remove the IP and everything works again. Later that night, added the IP back and everything craps out. Long story short, this is a bug in Exchange 2013 somewhere that prevents it from detecting the NIC dns settings when you add a secondary IP to the NIC. Quick fix is to specify the DNS settings in Exchange (I used my internal dns for External DNS and Internal DNS) as viewed in the screenshot above.
Other than that, I migrated Public Folders today and shutdown my Exchange 2010 servers…so far so good.