SQL Cluster instance network name - by SQLCraftsman

Status : 

  By Design<br /><br />
		The product team believes this item works according to its intended design.<br /><br />
		A more detailed explanation for the resolution of this particular item may have been provided in the comments section.

Sign in
to vote
ID 648682 Comments
Status Closed Workarounds
Type Bug Repros 0
Opened 3/1/2011 2:53:16 PM
Access Restriction Public


SQL Denali still uses the legacy naming standards of single-server named instances.  Given that each instance on a cluster has its own IP address and Network name resource, there is no reason to require a two-part name (virtualservername\instancename).  This legacy item is confusing, difficult to communicate to customers and users, and makes the product look unprofessional.  The Cluster installer and SQL Browser Service needs to allow the creation of single-part instance names for clustered instances.

This is particularly challenging when using DNS aliases to relocate SharePoint databases for disaster recovery purposes.
Sign in to post a comment.
Posted by Microsoft on 4/18/2011 at 11:13 PM

After triage we have denied this change request for SQL Denali.

It is true that current SQL Failover Clustering design is still depending on named instances, and unfortunately we cannot afford the risk to make fundamental change at the end of Denali development cycle.

We will consider this feature change in next major release. One specific issue need to be resolved is that current standalone instance listens on all the IP address available on its hosting box. Therefore if you have a standalone default instance side by side installed with a named clustered instance listening on port 1433 on the same node, you will have port conflict. Please be aware of this when you use the workaround provided previously in this thread.

Min He
Program Manager, SQL Engine
Posted by Allan Hirt on 4/6/2011 at 9:49 AM
Geoff is right here. Unlike multiple instances on a standalone machine, the names are always unique. In essence, you could have multiple default instances if you really wanted to. The problem is that isn't the way SQL Server is currently designed. I have the same relative issue when explaining naming on a cluster. What's worse is you can't rename the named instance part. For the most part, apps and clients can just connect with the Network Name. The problem comes in, say, when you want to configure replication. You're forced to use the full named instance.
Posted by Microsoft on 3/7/2011 at 9:26 AM
Dear customer,

If you would like to configure a clustered SQL Server instance to be referred to by only its cluster Network Name and to configure it so that you may use DNS CNAMEs to alias between target instances of SQL Server, you may configure the clustered instance to statically use TCP port 1433, after setup has completed.

This may be configured through the SQL Server Configuration Manager GUI, or programmatically through the SQL Server WMI Provider for Configuration Management.

Once you have configured the clustered instance to statically use TCP port 1433, then you may connect to the server using only its cluster Network Name or cluster IP Address, and then you may also use DNS CNAMEs to alias between target servers without any need for the two different target servers to share an instance name.

Alternatively, you may configure all your instances to use a single static port of your choice and access them using the connection string fragment "tcp:NetworkName,port". With that configuration, you could still use DNS CNAMEs to alias between target servers, since both instances would be using the same TCP port number on their respective host servers.
Posted by Microsoft on 3/3/2011 at 2:41 PM

Thank you for your feedback. We will evaluate this suggestion and contact you as appropriate.

Best regards,
Microsoft SQL Server