JDBC connection pool max-pool-size property in clustered environment
I have a question regarding the interpretation of JDBC connection pool's max-pool-size parameter.
Suppose we have a GlassFish domain consisting of 10 clusters, each having 2 instances. We also have a JDBC resource (say jdbc/myResource) associated with particular connection pool (say jdbc/myPool) and this resource is enabled on all clusters. The max-pool-size property of the connection-pool is set to 64. Now the question is: what's the maximum number of database connections that can be created by all instances in all clusters? Is it 64 or maybe 10*2*64 (1280)? If it's 64, then how does GlassFish enforces this limit in the cluster environment? Via DAS (what if it's down)? Some other way? And if it's 1280, then this definitely should be mentioned in the documentation.
I really hope I'll get the answer to that question.
Thank Y'all in advance.