Every time a passivate/activate!

Question

I've noticed some unexpected behaviour of stateful session EJBs in a
          WL6.0 cluster. Trace lines in the lifecycle methods reveal that when the
          ejb is first invoked the following sequence occurs (without any
          intervening time gaps):
          create / passivate / activate / run-the-method / passivate / activate
          and then subsequent method calls cause:
          run-the-method / passivate / activate
          The ejb is defined with replication type=in-memory,
          max-beans-in-cache=100 (only 1 in there during the test), and idle
          timeout seconds=3600.
          TIA
          Chris
          

Answer 1

You are correct except that the changes are replicated at the end of each
          transaction...
          Michael Reiche wrote:
          > makes sense to me... to implement replication, passivate is called to
          > replicate the changes to the secondary server whenever the EJB is modified
          > (maybe). Then activate is called to get the bean back to a working state.
          >
          > Mike
          >
          > Chris Palmer <[email protected]> wrote in message
          > news:[email protected]..
          > > I've noticed some unexpected behaviour of stateful session EJBs in a
          > > WL6.0 cluster. Trace lines in the lifecycle methods reveal that when the
          > > ejb is first invoked the following sequence occurs (without any
          > > intervening time gaps):
          > >
          > > create / passivate / activate / run-the-method / passivate / activate
          > >
          > > and then subsequent method calls cause:
          > >
          > > run-the-method / passivate / activate
          > >
          > > The ejb is defined with replication type=in-memory,
          > > max-beans-in-cache=100 (only 1 in there during the test), and idle
          > > timeout seconds=3600.
          > >
          > > TIA
          > > Chris
          > >