Hi All,
I'm wondering if I have configured something wrong or if this is expected behavior.
I have three nodes configured to all be
Windows Server 2012 R2
SQL Server 2016 Enterprise SP1
Availability Mode: Synchronous Commit
Fail over Mode: Automatic
Primary Role: Allow read/write connections
Readable Secondary: Read-intent Only
I have set up the servers to have a load balance read only (READ_ONLY_ROUTING_LIST=(('Node1','Node2'),'Node3))) etc etc
I'm running some tests to see how the system behaves before we go futher with the idea of using AG
I have two console applications, 1 that inserts one row every second, and another that reads the top row every second (using ApplicationIntent=ReadOnly) both are set to use the listener, no direct connection to the nodes are being created.
if the nodes that i'm hitting are either being promoted to primary or demoted to secondary (in a manual fail over), then the connections to these are lost for around 5-15 seconds
The node that is already a read only secondary and it not being promoted in the fail over remains connected and read requests continue.
Would you expect this to be the correct behavior for an AlwaysOn Cluster? or have I configured something incorrectly?