I have created an environment where 1 SAN is connected with 2 Nodes (part of SQL Server 2012 Failover Cluster) of SQL Server. Windows Server 2008 R2 is used, and single DB is stored in SAN.

We have achieved High Availability with 2 nodes and since there is a switch used to connect to SAN, it ensures that active node (which is one at a time) is connected to SAN, but other one is disconnected.

I wanted to know if I want to "scale" this architecture, how can we ensure that multiple "Active Nodes" are connected to SAN, is there a way somehow? This will help me add new nodes to SQL Server Cluster and try out various combinations like 2 active 1 passive etc.

I am from a developer background and not a server admin, please excuse if something is naive. Please help me understand. References to read are welcome.


First of all you need to be running SQL Server 2012 Enterprise edition.

Second, you need to create an AlwaysOn Availability Group(AAG).
You can have up to 4 replicas (5 total db servers) in an AAG.

You can change the AAG mode to synchronous or asynchronous. You also have the option of Read-Only replicas for things like Business Intelligence, Analysis, and Reporting.

This is one of the High-Availability configurations available in SQL Server 2012. You can combine a Failover Cluster with AAG.

  • I don't want to keep replicas, basically DB will have one "storage" copy in SAN. I don't want replication to come in pic. Our SAN will scale IO operations from multiple SQL Server Instances. I just want to connect Multiple SQL instance to provide scalability, which currently is not happening due to Switch. AAG's create replicas as far as I know. Please excuse if I got you wrong. – Ashish Jain Feb 7 '14 at 8:49
  • I see what you mean. Wouldn't using a SAN for this be redundant, since the SAN priorities IO over two paths? AAG does create replicas. – Vasili Syrakis Feb 7 '14 at 9:01
  • Two paths? Basically it should point to "same file system" single copy of DB, its kind of concurrent access to same Filegroup via multiple instances of SQL Server. I think it would be difficult to achieve in SQL Server. – Ashish Jain Feb 7 '14 at 9:08
  • serverfault.com/questions/450478/… Shows my approach is wrong. Thanks for help. – Ashish Jain Feb 7 '14 at 9:15

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.