SQL Server – How to Query Readonly Replica in Always On Cluster

availability-groupsdistributed-transactionssql server

We have an Always On "cluster" that consists of 2 servers (there will be more), so one is PRIMARY and other(s) is(are) SECONDARY. The idea was to dedicate SECONDARY as a readonly replica, so it would be a somewhat search-server.

I do know how to set ConnectionString in the C# so it will use SECONDARY replica – just add ApplicationIntent=ReadOnly.

But I don't understand how to address the SECONDARY from stored procedure if I'm using distributed query. Is there a way to set some parameters to the query (like WITH statements or something), so query would use only replica and not the PRIMARY node?

The thing is, obviously from the start server1 is PRIMARY and server2 is SECONDARY, but when server1 fails then server2 is PRIMARY and server1 will be SECONDARY after it get fixed. So I cannot just use the static server2's name.

So far I've managed to get current replica's server name as a variable and use it in EXEC:

-- find first available replica
DECLARE @replicaServer nvarchar(50)

SELECT TOP 1
    @replicaServer = RCS.replica_server_name
FROM
    sys.availability_groups_cluster AS AGC
    INNER JOIN sys.dm_hadr_availability_replica_cluster_states AS RCS
          ON RCS.group_id = AGC.group_id
    INNER JOIN sys.dm_hadr_availability_replica_states AS ARS
          ON ARS.replica_id = RCS.replica_id
    INNER JOIN sys.availability_group_listeners AS AGL
          ON AGL.group_id = ARS.group_id
WHERE ARS.role_desc = 'SECONDARY'
      AND connected_state = 1

-- and query it
DECLARE @cmd nvarchar(MAX) = 'SELECT * FROM [' + @replicaServer + '].[somebase].[someschema].[sometable]'
EXEC (@cmd)

But that's a shame, I guess.

Best Answer

After some help and hints (especially thanks to Andriy M) I've managed to understand what I actually want and how to do it. This article from Microsoft helped as well: https://technet.microsoft.com/en-us/library/hh403414.aspx

First I had to create a read only listener. And then at the other server add linked server like this (borrowed from: https://stackoverflow.com/a/19585304):

USE [master]
GO
EXEC master.dbo.sp_addlinkedserver
    @server     = N'cluster-readonly', -- just a name
    @srvproduct = N'SQL', -- anything except 'SQL Server'
    @provider   = N'SQLNCLI11',
    @datasrc    = N'ClusterListener', -- name of the readonly listener
    @provstr    = N'ApplicationIntent=ReadOnly',
    @catalog    = N'AnyActualDataBase' -- system database like 'master' is no good
GO

How do I check, where am I:

SELECT * FROM OPENQUERY([cluster-readonly], 'SELECT @@SERVERNAME')

Here's some trick - if I don't specify @catalog, then connection will be to the PRIMARY node for some reasons.

Related Question