2 Replies Latest reply on Dec 5, 2012 10:15 AM by andy.miller

    Bean Instance Pools

    andy.miller

      During the development meeting in Brno in November, I was told that the bean instance pools were specific to each bean.  For example, if you had two stateless session beans, let's just call them Bean A, and Bean B, you could define a strict max instance pool for each one.  For the life of me I don't see how you can actually do this with the current schema.

       

      There is nothing in the schema to associate a bean, by name or any other attribute, to a pool.

       

      Is there a way to do this?

       

      How does this really work?  It appears that there is really just one pool, and everything shares it, even if there are multiple deployments.

        • 1. Re: Bean Instance Pools
          jaikiran

          The (default) pool configurations that you see in the standalone/domain xmls is the pool "configurations" from which a per bean (runtime) pool is constructed. This is the default pool that gets applied to each of the beans, unless the deployment or the bean itself overrides the pool name to use. Overriding the pool name for each bean or each deployment is done via deployment specific descriptors or annotations. More details about it are here https://community.jboss.org/thread/175628

          • 2. Re: Bean Instance Pools
            andy.miller

            Thanks Jaikiran.  This now makes sense.  Now I can setup a test to see whether one pool shared by many beans is better than multiple pools or a pool for each bean.