RE: https://msdn.microsoft.com/en-us/library/ff878487(v=sql.110).aspx#DotNetHotfixes

Personally I Struggle with A WSFC node can host only one availability replica for a given availability group = one or more instances of SQL Server can host availability replicas for many availability gr oups and ive been unable to locate a decent online explanation of this Statement.

My client wants to Install 3 SQL Instances on a two node Syncronous mode cluster with 1 AG in Each SQL Instance ,  Ie three sql instances on Each Node and has been told this is supported I struggle with the Failover being at node Level and the above statement.  ie If 1 AG Group Fails over the physical node what state are my other 2 instances in? How Do I nswer the above question "how many WSFC nodes   are required for to support the availability replicas of the planned   availability groups." ?

Regards

Hi RobinsonColin,

As your post, the design that installing three SQL Server instances on each node and creating an AlwaysOn Avaliability Group (AG) per instance is supported. You can create different listeners for each AG, each AG can failover independently without impacting other AGs. What we need to note is that ports configuration of involved SQL Server instances and listeners should be proper, for more details, please review this similar thread.

In addition, a WSFC node can host only one availability replica for a given availability group, and
an AlwaysOn Availability Group supports up to eight secondary replicas with SQL Server 2014. For example, if you configure a synchronous secondary replica and an asynchronous secondary replica for an AlwaysOn Availability Group, you need to have three WSFC nodes.

Thanks,
Lydia Zhang

If you have any feedback on our support, please click here.

and the other part of my Question ?

" If 1 AG Group Fails over the physical node what state are my other 2 instances in?"

Colin

and the other part of my Question ?

" If 1 AG Group Fails over the physical node what state are my other 2 instances in?"

Colin

yep,  Two physical servers each hosting 3 Instances

server1\Instance1\Ag fails to server2\Instance1\AG   (two node synchronous)

server1\Instance2\Ag fails to server2\Instance2\AG   (two node synchronous)

I hope this is a clearer view of my requirements

Server1\Instance3\Ag Fails to Server2\instance2\AG  (two node synchronous)

Three possible Scenarios:

1.  sql Service Fails on a single instance

2. A single Network card Fails on server1, Secondary Network is available, Cluster Fails over the Physical Replica (ie server1 at the Node level).

3: Total Hardware Failure of Server1. 

What is the Supported Failover State in each of these Scenarios

finally an admission from MS that this multi Instance configuration does not support automatic failover. manual intervention is required.  Obviously that manual intervention depends on the nature of the Failure.

However the link provided does not provide any further details. Microsoft needs to Fully document this scenario in relation to the initial Paragraph that headlines this post making it clear that multi instance to Multi Instance is not FULLY supported.