3

In clustering you specify DNS name/IP address of cluster that is then used by all applications. In case of failover, the same DNS name/IP address will point to the appropriate database server.

In Always on Availability Groups, there is concept of Listener which to my understanding does the same as above, so do we really need a Listener when the Failover clustering already handles this concept (IP address) of failover?

2

do we really need a Listener when the Failover clustering already handles this concept (IP address) of failover?

A SQL Server Failover Cluster Instance (FCI) will always have a "Network name resource" for clients to connect to the active node.

But with AlwaysOn Availability Groups it is optional, and created through SQL Server. If you don't create an AG Listener, the Windows cluster will not have any IP address or hostname that moves along with the AG.

In any case an AG Listener is a Windows Server Failover Clustering "Network name resource", and so is confined to a single Cluster. If you have a Distributed Availability Group which spans multiple Clusters, or a Read-Scale Availability Group which has no cluster, then you can't have an AG Listener. (Although with a Distributed AG each participating AG can have its own listener).

  • Is Listener a concept of Clustering or Availability Groups? – variable Oct 10 '18 at 5:06
  • Please can you also mention about situation where we have a Cluster C1 (with 2 nodes N1 and N2) and AG is setup between this cluster C1 and another SQL server instance N3. How does the FCI and Listener behave in this scenario please? – variable Oct 10 '18 at 5:12
  • Hi, any information/guidance/reference/links will be very helpful please – variable Oct 10 '18 at 11:43
  • See update to answer. – David Browne - Microsoft Oct 10 '18 at 13:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.