After several calls with Microsoft, a ROSS, and several other engagements, we have the following configuration below.
Goal
Each DAG (Exchange 2013, and 2010) has 4 members, two in each site. Assume that the cluster resource is in the respective "active sites". Public folders are on Exchange 2010/Legacy
Clients
Outlook 2007, cached and non cached mode and newer. Some are on-network, some are off network
New York Subnet: 10.10.15.x
Autodiscover: autodiscover.contoso.com
DAG Name: NYCDAG , No Cluster Administrative point
RPC setting on all Exchange 2013 databases: NameOfA2013Server.contoso.com
RPC setting on all Exchange 2010 databases: old2010RPC.contoso.com
ClientAccessArray: old2010RPC.contoso.com
Internal URLs and External URLs for all Virtual Directories (2013 servers)
https://NYCCAS.contoso.com/x (x is OAB, EWS, OWA,
Note- we did have an ambiguous namespace where Autodiscover URL, and the RPC name were the same. This has been corrected.
Public Folder server PF01: OutlookAnywhere is enabled, and has a name of PublicFoldersNYC.contoso.com. 1:1 NAT available to outside
Exchange 2010 CAS Array: OutlookAnywhere is enabled and has a name of Legacy2010.contoso.com. 1:1 NAT available to outside.
San Francisco: 10.5.5.X
Autodiscover: autodiscover.contoso.com
DAG Name: SFDAG (stretched DAG with NYC). No Cluster Administrative point
Internal URLs and External URLs for all Virtual Directories
https://SFCAS.contoso.com/xxx
2010 Public folders, CAS not yet configured.
Is it strange that we have 3 OutlookAnywhere endpoints in the NYC site? What is the preferred configuration?