- From: Jim Ma <mail2jimma@gmail.com>
- Date: Thu, 14 Jun 2012 12:54:32 +0800
- To: Doug Davis <dug@us.ibm.com>
- Cc: public-ws-addressing@w3.org
- Message-ID: <CAAvMoHmMNROvpn0X6BVp09B_QKdKPaXrgXLVXZumf-XApT9DnQ@mail.gmail.com>
Thanks for the link, Doug. On Wed, Jun 13, 2012 at 8:17 PM, Doug Davis <dug@us.ibm.com> wrote: > > If you look at: http://ws-i.org/Profiles/BasicProfile-2.0-2010-11-09.html section 3.7.10, and in particular R1148, it might provide the guidance > you're looking for. > > thanks > -Doug > ______________________________________________________ > STSM | Standards Architect | IBM Software Group > (919) 254-6905 | IBM 444-6905 | dug@us.ibm.com > The more I'm around some people, the more I like my dog. > > > *Jim Ma <mail2jimma@gmail.com>* > > 06/12/2012 06:34 AM > Please respond to > mail2jimma@gmail.com > > To > public-ws-addressing@w3.org > cc > Subject > WSA FaultTo Question > > > > > Hi All, > When I looked at a wsa faulto issue, I didn't find there is > some specification conformance to address how to > respond to the client when the fault occurred in the following ws > invocation scenario: > > 1. A client sends a request to invoke a request/response operation from > EndpointA, the request's wsa faultTo header point another endpointB > and not the requester > > 2. There is error occurred during EndpointA processing the request, so > the fault message is sent to EndpointB to notify the faultTo target > endpointB which is specified in request wsa headers. > > 3. It's a two way webservice invocation , so we need to reply to the > initial requester. > > So my question is what should the endpoint respond to client/requester? A > 500 > error with the same fault message sent to faultTo endpoint or we need to > forward the response from the faultTo target EndpointB to requester ? > Or EndpointA should > mix this message and let the requester clearly know the exact soap fault > message, and also if the fault message has successfully notified the > endpointB ? Which option do you think > it's the ideal way to handle this ? Is there any future specification > will address this scenario ? > > Thanks in advance ! > > Jim > > > >
Received on Thursday, 14 June 2012 04:55:02 UTC