Hi,

We have a 3 node cluster running Windows Server 2012 R2 with 5 SQL clustered roles running SQL 2012 spread across them. For certain reasons we have a requirement to install a second SQL named instance onto one of the existing SQL roles. All of our SQL instances are currently running on different ports with no issues.

When we run the wizard to install the additional instance we specify the new instance name along with the existing network name to install to. However, we receive this error message:

"The SQL Server failover cluster instance name <existing role name> already exists as a clustered resource. Specify a different failover cluster instance name."

I am wondering if it is by design that it doesn't allow you to do this or a limitation of the GUI install? Is running multiple SQL named instances on a single cluster resource group/role supported?

Cheers
Brady

• Edited by Kenman87 Wednesday, March 11, 2015 3:27 AM

Hi Brady,

"The SQL Server failover cluster instance name <existing role name> already exists as a clustered resource. Specify a different failover cluster instance name."

As other post, you can run multiple named instances on a cluster.

When adding SQL Server instance on the cluster, SQL Server installer will actually checks whether input SQLNetworkName resource is already there in the cluster, if it finds any orphan resource related then above error will be displayed. In this case, we can delete the orphaned resource following the steps in this blog.

However, in your scenario, to bypass the error, you can simply specify a different SQL Server Network Name when installing the second instance.

Thanks,
Lydia Zhang

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

Thanks for the info guys, however I just want to clarify what my situation is as I don't think we are quite taking about the same thing.

Say for example I already have a cluster group, with a network name, called "clusterrole1". I've already installed a SQL instance into this cluster group called "inst01". So essentially in order to connect to that SQL instance my connection string would be clusterrole1\inst01.

My question was related as to whether I could install a second SQL instance, for example "inst02" into the existing cluster role. That means I would have a cluster role/resource group with 2 SQL instances within it.

ie:
clusterrole1\inst01
clusterrole1\inst02

Both instances would be configured to listen on different ports so I don't see why this wouldn't work?? What would be technically stopping me from installing the second instance. Not much different from having a standalonee server with 2 SQL instances installed on it which is defiantly possible.

From a clustering perspective is there a technical limitation to doing this?

Cheers
Brady

Hi Brady,

Based on my understanding, in the Instance Configuration page as below, you specify the same SQL Server Network Name of the first SQL Server instance when installing the second SQL Server instance , right?



If that is the case, we need to specify a different SQL Server Network Name for the second SQL Server instance. As Seans post, these things (including instance name, SQL Server Network Name and so on ) must be distinct  in cluster, the behavior is designed to block someone from creating resources that would conflict.

Please review the detailed difference of installing first & second SQL Server  instances from below blogs.
SQL 2012 Failover Cluster Build (Step by Step) Part 1: Installing SQL 2012
SQL 2012 Failover Cluster Build (Step by Step) Part 3:Adding a New Instance

Thanks,
Lydia Zhang

                                                                                                            

Hello,

Say for example I already have a cluster group, with a network name, called "clusterrole1". I've already installed a SQL instance into this cluster group called "inst01". So essentially in order to connect to that SQL instance my connection string would be clusterrole1\inst01.

Incorrect. You have a resource group named ClusterRole1, a named instances where the instance part is named Inst01, but we have no idea what the virtual computer object (VCO) is, and that would be what makes up your name. There will be a network Name in the resource group, your connection will be NetworkName\Instance unless it's the default instance, then it would just be NetworkName.

The resource group can be named whatever it wants to be - that doesn't have any effect on SQL Server in a cluster, it's just the name for the collection of resource in the cluster and isn't externally facing or used anywhere except inside the cluster.

My question was related as to whether I could install a second SQL instance, for example "inst02" into the existing cluster role. That means I would have a cluster role/resource group with 2 SQL instances within it.

This will not work. period. You'd need a separate resource group to install it into (the install will create it), IPs, VNN, VCO, Disk, etc.

From a clustering perspective is there a technical limitation to doing this?

The limitation is the implementation and rational of doing this with SQL Server. Setup checks for these things and errors appropriately. It doesn't make sense to have two instances in the same resource group for clustering. Cluster is used for HA, so having a single instance take down all instances doesn't make sense from that point of view alone, not to mention other potential issues that could occur.

If you want to do this, run the installer again and create another clustered instance of SQL Server in its own resource group.

-Sean

No worries! Thanks to both of you.