> Thank you for sending this along. Looking at the sample, it does not > appear that anyone is listening on port 9990 (the callback). Am I missing anything?
Do you mean, an endpoint isn't explicitly created by the client application? There's no need for this, as it should be created under the hood by the HTTP conduit as a result of the decoupled response endpoint config. > Also.. the client code sleeps for 30 seconds. That's just because the demo simulates message loss by throwing every other outgoing message on the floor (via an interceptor). The delay just allows the resend logic to kick in. > This may mask the issue.. since my client is short lived (test client). So.. the client worked and then goes away. Is your client explicitly shuting down the Bus? This is required so that the client-side WS-RM layer can gracefully terminate the message sequence (forcing any pending ACKs to be sent). Cheers, Eoghan 2009/3/19 Bryan Kearney <[email protected]> > Eoghan Glynn wrote: > >> Do you have an example on the website? >>> >> >> Have a look at the ws_rm sample, particularly the client-side config: >> >> >> http://svn.apache.org/viewvc/cxf/trunk/distribution/src/main/release/samples/ws_rm/src/demo/ws_rm/client/ws_rm.xml?view=markup >> > > Thank you for sending this along. Looking at the sample, it does not appear > that anyone is listening on port 9990 (the callback). Am I missing anything? > > Also.. the client code sleeps for 30 seconds. This may mask the issue.. > since my client is short lived (test client). So.. the client worked and > then goes away. > > -- bk > >
