[BYTEIO-WG] Error with Genesis II Client talking to Unicore

[Mohammad Shahbaz Memon]

Hi Mark,

Thanks for identfying these problems.

append/truncAppend problem: unfortunately append/truncAppend
operations return messages are missing with transfer mechanism
information.

seekRead/seekWrite action URL: Your guess was right :), indeed its a
typo problem. I just checked ,it was seek-read / seek-write instead of
what is stated in WSRF rendering document.

UNICORE end point is down till tomorrow. This will be up by tomorrow
with the improvement of aforementioned problems.

Regards,
Shahbaz



Thanks for identifying core compliance problem

On 7/26/07, Mark Morgan <mmm2a at virginia.edu> wrote:
> Shahbaz,
>
> Thanks for the info.  We've changed out code to send out the actions
> (which are required by WS-Addressing, but I believe required to be
> ignored by WS-I).  At any rate, we are sending them out and they seem to
> work for most of the test cases.  We can now successfully communicate
> with your service for tests 4.1 - 4.7, and 5.1.  However, we fail to
> interoperate with your service endpoint on the append/truncAppend
> operations, and on the seekRead/seekWrite operations.  For the
> append/truncAppend operations we are failing because the response that
> we get back from your service contains an empty appendResult data
> structure.  However, according to the ByteIO WSRF Rendering
> documentation, this result element must contain a transfer-information
> element, even though there is no data to be transferred for the
> operation.  Please see attached file Gen2UNI.append.log for an example
> session showing this issue.
>
> For the seekRead/seekWrite operations we are getting a fault back
> indicating that the SOAP Action in the header is not understood.  As you
> can see from the attached session log (Gen2UNI.seekRead.log), I think we
> are passing the SOAP action specified by the WSRF Rendering document.
> As I stated above, technically I'm not sure that you are allowed to
> route messages based on this action anyways, but even if you are, I
> suspect that you may have a typo in the action that you are looking for
> (perhaps you are still using the random namespace in those actions?).
>
> -Mark and Karolina
>
> On Thu, 2007-07-26 at 21:31 +0200, Mohammad Shahbaz Memon wrote:
> > Thanks for pointing out this problem.
> >
> > The soap message you sent has actually no wsa:Action element.
> >
> > In our WSRF implementation, operation on any ws-resource has to
> > specify this element.
> >
> > Shahbaz
> >
> >
> >
> > On 7/26/07, Mark Morgan <mmm2a at virginia.edu> wrote:
> > > The good news is that the changes suggested yesterday have fixed the
> > > problem that we had with creating new resources on the UNICORE interop
> > > factory.  Unfortunately, now that that is working, we seem to be having
> > > a problem with communicating with that new endpoint.  We are getting a
> > > response back from all operations on the RByteIO Endpoint indicating
> > > that the WS-Addressing headers are not being parsed correctly.  A
> > > cursory look at the messages that we are sending don't reveal any
> > > obvious problems.  Please find attached an example session including a
> > > successful createResource/createResourceResponse pair of messages, and a
> > > failed GetResourceProperty/BaseFault pair of messages indicating the
> > > aforementioned problem.
> > >
> > > -Mark and Karolina
> > >
> > > --
> > >   byteio-wg mailing list
> > >   byteio-wg at ogf.org
> > >   http://www.ogf.org/mailman/listinfo/byteio-wg
> > >
> > >
> > >
> >
> >
>
>


-- 
------------------------------------------------------------------------
Mohammad Shahbaz Memon
Department of Applied Mathematics
Forschungszentrum Jülich GmbH
Jülich Germany

Skype: m.memon
Web page: http://www.fz-juelich.de/zam/ZAMPeople/memon_m
Office: +49 (0)2461 61 6567
Fax:     +49 (0)2461 61 6656