I would like confirmation on our new Exchange 2010 design that is in production. It is regarding the CASArray.
Our 1 DAG design is as follows:
AD Site DC1 (Active)
3 Hub/CAS servers (DC1CASARRAY.company.local)
3 Mailbox Servers
AD Site DC2 (Passive)
2 Hub/CAS servers (DC2CASARRAY.company.local)
2 Mailbox Servers
All users are located in DC1. DC2 is activated in the event of a partial or total site failure.
Here is my scenario/question.
We had a DC1 failure recently, and I activated all databases in DC2. ALL databases have DC1CASARRAY.company.local as their rpcclientaccessserver attribute.
The DC1CASARRAY DNS record points to a VIP on our load balancer in DC1. There is currently no record for DC2CASARRAY in DNS.
With this setup, I technically do not need a DNS record for DC2CASARRAY correct? Since all the databases have the DC1CASARRAY attribute, all I would need to do in the event of a site failure, is to change the DC1CASARRAY DNS record to point to the load balancer VIP in DC2 (which essentially would be the DC2CASARRAY DNS Record).
It seems like in this setup, DC2CASARRAY isn't really necessary (unless in the future we would add active databases in DC2, add another DAG, etc).
Now - back to the fail over event. This was the first failure since upgrading to 2010, so I'm fairly new at the process (3 months). In the midst of the fail over, i actually ADDED a DNS record for DC2CASARRAY to point to the VIP in the DC2 site. Looking back on it, I should have just changed the DC1CASARRAY record to point to the VIP in DC2. Weird thing is, it worked, and I'm trying to understand how. Clients connected to DC2CASARRAY (validated by ctrl+clicking Outlook icon, connection status) and were able to get mail.
We've since failed back and I deleted the DC2CASARRAY record, and clients connected to DC1CASARRAY just fine.
Any assistance to help me understand is appreciated!