SQL Server: Connect to a Named SQL Failover Cluster Instance (FCI)
Commonly seen today is this binding order used to connect with a named CLUSTER INSTANCE i.e. SQLcap + Ins Nam + Port.
SQL’s CAP | Instance Name | Port | Connectoid | While this works |
SQLa | Default | M | SQLa | SQLa |
SQLb | Xx | N | SQLb\Xx,N | SQLb |
SQLc | Yy | O | SQLc\Yy,O | SQLc |
This can be set to work with SQLcap only, after changing the PORT to 1433.
Benefits
Users get rid of
- A mumbo-jumbo (SQLcap + Ins Nam + Port) and uses an easier to remember token.
- Maintaining a connection string with (SQLcap + Ins Nam + Port).
- Creating aliases in case of migration or planed exodus.
- Easier renames.
- And we can delete the browser service DLL to save disk space (joking).
Effort Needed
- Amend the SPN creation documentation of yours.
- Update support guides that [SQLcap] is actually this [SQLcap + Ins Nam + Port].
So, token is not actually what you get from select @@servername (which remains SQLcap + Ins Nam). This works well with…
- SSMS, OSQL, SQLcmd
- UDL
- SQL Server Native Client.
- MS OLE DB Provider for SQL Server.
Now this is worthwhile for fat setups (multiple SQL FCI consolidated on a single WSFC).
Kindly share your thoughts if you’ve done this in the past or any good/bad experience around this.