I've just read Brian Knight's articles on setting up A/A Clustering and I'm slightly confused about the naming of instances.
As I understand it, in an A/A configuration SQL Server will be running on both nodes (A and B).
On Node A, I would create a new virtual server (VirtA) with a default instance.
On Node B, I would create another virtual server (VirtB). However this requires a named instance. Why can't this new virtual server have a default instance as the connection from the client will use the Virtual Server name, i.e. Client connects using VirtA or VirtB. From what I've read the Client would need to connect using VirtA or VirtB\InstanceName.
many thanks, Justin
I'm not sure I understand the question completly. One, if you run an A/A setup you are really running two sql servers that can fail over to each other. Thats why you need seprate named instances. If I was writing an app to pull from a group of sql servers it would determine which one to pull from. Actually both instances are named virtual instances. When you run the setup on the primary node you will see default instance is grayed out.
Please note the following , Brian speaks in General through this article and note further whether its A/A or A/P , the client will never know which SQL SERVER he is connected to as it is not important , since it is a DB availability issue.
Also note that in any one of these a default instance is not present as a Quorum Disk takes care of instances.
I hope you have got it?
Viewing 3 posts - 1 through 2 (of 2 total)