I am trying to google/bing my way through this, but here are the main points:
- All remote Exchange servers had a) static IP addresses and b) were instructed to use DNS servers in Edmonton and Calgary
- If WAN links dropped, they lost DNS
- SOLUTION! Point remote servers to the domain controller in that location as the primary DNS server. WAN outage means Exchange still resolves names locally.
Well the day came and we add the local DNS server for each remote site as the primary DNS server at the top of "DNS server addresses, in order of use" portion of advanced TCP/IP settings dialogue box.
- A reboot is performed
- Voila! Exchange shits the bed! Hub transport servers in Edmonton and Calgary cannot resolve remote server names; as in no host(a) record exists for them anymore. And we have a 91 minute outage until the DNS is sorted out (ipconfig /registerdns run on each server).
Has anyone encountered this or have some deep knowledge of DNS (relative to my own) that could at least throw out a theory as to why this might happen?
- Edited by Kelly Shields 10 hours 51 minutes ago