I have a 4 node AlwaysOn configuration with 2 nodes at the main site, 2 at the DR site and a FSW. The 2 main site nodes are configured for Automatic/Synchronous fail-over, the 2 DR nodes are Manual/Asynchronous and transnational replication from one DB is configured to a remote subscriber. I configured my replication following this MSDN article and tested that replication works fine: https://msdn.microsoft.com/en-us/library/hh710046(v=sql.120)
My issue is with fail-over to the DR site to simulate the 2 primary site nodes being unavailable but not completely lost. The 2 DR nodes come online and I can resume data movement between them without issue. Changes to the configured publication DB also synchronize between the 2 DR nodes but nothing is delivered via replication to the subscriber.
I'm looking for a suggestion to get replication working without ejecting the 2 primary site nodes from the Availability Group. The intent of the exercise would be to fail back to either of the main site nodes when they are available again, but not delay replication while the primary site nodes are offline.