Folks: I know this is probably a dumb question (I'm a relative newbie), but why does the "javax.naming.InitialContext" method participate in transactions? Here's the scenario: - I have jboss 2.0 Final running in one VM, and 2 client applications (each in different VM's) attaching to it. - All beans (and all methods) have transaction attribute set to "REQUIRED". - The first client app calls a stateless session bean method. - While this first app is running this method, the 2nd client app calls "new InitialContext()" and then "context.lookup( ejbname )", on an entity bean. - The call to "lookup" will hang/lock until the first app's bean method completes! Once the session bean method returns, the JNDI lookup completes successfully. (Note: I can see this behavior very explicitly with my source level debugger, where I stop the first client at the first statement in the session bean method, before it's actually done anything, and then try to step the 2nd app past the "lookup" statement.) Thus, everything works, I just don't understand why there's synchronization/locking on the JNDI call. Am I crazy? :-) If I'm right, this seems an exorbitant performance problem; surely I must have something set up inefficiently? Help! Thanks, Roger -- -------------------------------------------------------------- To subscribe: [EMAIL PROTECTED] To unsubscribe: [EMAIL PROTECTED] List Help?: [EMAIL PROTECTED]
