Anatoly Akkerman wrote: > David Jencks wrote: > >>4) invocation is routed to correct transport mechanism (such as by the HA > >> Invoker interceptor)(AFAIK not yet written as an interceptor) > >>5) Invocation is received at server side. > >>6) The server decides if the resource registration callback is needed. > > > > And the resource registration occurs on the return of the original method > > invocation, or does it take an extra round trip immediately? The former > > incurs redundant bookkeeping requirements and the latter is inefficient. > > > The latter is completely out of the question performance-wise, so I hope > Ole is not argueing for it.
I'm trying to argue that the latter can be avoided without any client-side changes, if the tx is not needed on the server side. > It seems that sending the resource registration request on the return of > the original method invocation should not be any more complex than what > David does right now. Please consider recovery after failures: If the registration is delayed until the call returns, a client crash during the call would leave an orphaned tx branch at the server, with no recovery possible. It might be better to do the registration before the call, like David currently does. Best Regards, Ole Husgaard. ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Jboss-development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development