A simple query about the data source. I have setup a data source in a cluster with 2 or more managed servers. The data source in turn points to a connection pool. Users acces the site via a load balancer. When a request comes in from the load balancer, does the data source provide a connection pool of the server it is bound to, or does it load balance the request.

An example would be - say each connection pool (Cp1, Cp2, Cp3) is setup for 5 connections. Then say that the first connection pool (Cp1) is fully occupied and a request comes in to the data source, will the data source provide a connection from another connection (Cp2,Cp3) or will it simply queue up the request until it can get a connection from Cp1 ?