perfsonar development work

[Loukik Kudarimoti <>]

Matthias K. Hamm wrote:
> Hi Loukik, all,
>
> Loukik Kudarimoti schrieb:
> > Hi Michalis, Matthias and Mark  (Szymon and Nicolas, please see the 
> > ending of my reply)
> >
> > Please correct me if I am wrong,
> > 1) The latest RC is now using document/literal style of web services 
> > just like all other perfSONAR services.
> Yes.
> > 2) Also, only one request is supported (the one that you mentioned)
> Indeed.
> > 3) for anything else, the service returns soap exception (no error code)
> That`s correct.
> > 4) Further, even within the supported requested, the service 
> > sometimes ignores bits which need to be taken care of (such as 
> > missing metadata reference or metadataId).
> Our MP does not make use of any metadata references. Perhaps we should 
> throw an error if we receive such information (better than ignoring 
> it) ...?
> > 5) LS registration is not supported yet (expected in the next RC)
> Ack.
> > 6) (new) Echo request not supported
> No.
>
> Uups... quite a few unexpected additional things to implement, test 
> and document. Let's talk about the timeline again:
>
> - We are currently quite busy with bug searching/fixing and proper 
> documentation for the 2.0 release of the central part of the 
> monitoring system.
> - We can try to implement (3)/(4) for E2Emon MP until end of July. Is 
> it possible to add the MP to the perfSONAR release at this stage?
> - Due to our holidays we have a "reduced workpower" in 
> August/September, so we can promise you the next micro release 
> including (5) and possibly (6)  not before end of August.
>
> What do you think about?
It is possible to add the next micro-release with (3) and (4) into the 
next update.

Nicolas was recently talking about the 2.2 update. There will surely be 
one coming in future - maybe september/october time frame. We can then 
update the E2EMon with (5) and (6) features.

This sounds a good plan to me.

Loukik.

> Cheers,
>
> M&M
> > I think having (1) & (2) is already a good progress for this service. 
> > We would definitely like to have (3) and (4) as well.
> >
> > Basically, with (3), the service converts all internal exceptions 
> > into an xml error code specification. (see page 68 of this document: 
> > http://anonsvn.internet2.edu/svn/perfsonar/tags/JAVA-SQL-MA-1.3/doc/ma/sqlTypeMA/Interface_Specification.doc 
> > .  For an example, see page 14, Example 7 in the same document).
> >
> > Result codes / error codes should definitely be used whenever your 
> > service cannot do a normal operation (such as malformed request, 
> > unsupported request, system configuration problems, missing files, 
> > missing data bits, etc).
> >
> > With (4), its down to "screw tightening" inside your code (i.e., 
> > doing a thorough check on the request).
> >
> > I understand that (5) is for the next micro-release.
> >
> > With (6), we might have discussed about echo/ping functionality 
> > before. Its like an application level ping where a particular xml 
> > message called echo request is sent. The service, when it reads the 
> > message, responds with an echo response message. Both messages are in 
> > xml).
> > I think its very beneficial to have this functionality and is very 
> > easy to implement it (It shouldn't take more than a couple of hours). 
> > But I would say that its up to you guys. All services in perfsonar 
> > 2.0 (except bwctl mp) support this functionality.
> >
> >
> > As far as a "definition of what is a proper perfSONAR service" is 
> > concerned, we don't have any such document yet.
> > Nicolas, Szymon, any thoughts?
> >
> > Loukik.
> >
> >
> > Matthias K. Hamm wrote:
> > > Hi Michael,
> > >
> > > Michael Michalis schrieb:
> > > > Hi to all,
> > > >
> > > > I have just finished testing the E2EMon MP.
> > > fine :-)
> > > > Installation went OK although when the service was called this 
> > > > appeared on the screen "Use of uninitialized value in join or 
> > > > string at /usr/lib/perl5/vendor_perl/5.8.5/SOAP/Lite.pm line 1106."
> > > >
> > > > But there seem to be some serious problems regarding the service 
> > > > since the service has failed all the tests but one. The only 
> > > > request that the service responded correctly was the valid and not 
> > > > malformed request  like this one:
> > > >
> > > > <nmwg:message type="SetupDataRequest"
> > > > xmlns:nmwg="http://ggf.org/ns/nmwg/base/2.0/";>
> > > > <nmwg:metadata id="meta1">
> > > > <nmwg:eventType>Path.Status</nmwg:eventType>
> > > > </nmwg:metadata>
> > > > <nmwg:data id="data1" metadataIdRef="meta1"/>
> > > > </nmwg:message>
> > > That's what it is supposed to do ;-)
> > > > All of the other tests involving malformed or invalid values of 
> > > > necessary parts (as described in the service setupDataRequest rnc 
> > > > file) have failed.
> > > > The service seems to be unable to produce error result codes as it 
> > > > is accustomed by other perfSONAR services like this:
> > > >
> > > > <nmwg:metadata id="*resultCodeMetadata*">
> > > >  <nmwg:eventType>warning.ma.metadata_configuration</nmwg:eventType>
> > > >  </nmwg:metadata>
> > > > - <#> <nmwg:data 
> > > > id="*resultDescriptionData_for_resultCodeMetadata*" 
> > > > metadataIdRef="*resultCodeMetadata*">
> > > >  <nmwgr:datum 
> > > > xmlns:nmwgr="*http://ggf.org/ns/nmwg/result/2.0/*";>MetadataConfigurationStorageManager.fetch: 
> > > > No metadata in the metadata configuration</nmwgr:datum>
> > > >  </nmwg:data>
> > > >
> > > > including an eventType element and a datum element describing the 
> > > > error. In fact as far as i can tell that kind of behavior is not 
> > > > documented in the setupDataResponse rnc file at all.
> > > >
> > > > The service when not responding with a  SOAP Fault it responds with 
> > > > an error message somewhere in the soap envelope and not inside the 
> > > > setuDataResponse as it should.
> > > >
> > > > <soap:Envelope 
> > > > xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"; 
> > > > xmlns:xsd="http://www.w3.org/2001/XMLSchema"; 
> > > > xmlns:soap="http://schemas.xmlsoap.org/soap/envelope/";>
> > > >   <soap:Body>
> > > >      <s-gensym73 xsi:type="xsd:anyURI">
> > > >         ERROR: bad type of request, eventType Status not understood 
> > > > by this MP (
> > > >         <soapenv:Envelope 
> > > > xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/";>
> > > >            <soapenv:Header/>
> > > >            <soapenv:Body>
> > > >               <nmwg:message type="SetupDataRequest" 
> > > > xmlns:nmwg="http://ggf.org/ns/nmwg/base/2.0/";>
> > > >                  <nmwg:metadata id="meta1">
> > > >                     <nmwg:eventType>Status</nmwg:eventType>
> > > >                  </nmwg:metadata>
> > > >                  <nmwg:data id="data1" metadataIdRef="meta1"/>
> > > >               </nmwg:message>
> > > >            </soapenv:Body>
> > > >         </soapenv:Envelope>
> > > >         )
> > > >      </s-gensym73>
> > > >   </soap:Body>
> > > > </soap:Envelope>
> > > >
> > > >
> > > > Further more  when the metadataId Ref attribute of the request is 
> > > > not present or is invalid the service still responds as normal(In 
> > > > the rnc file the id attribute is considered mandatory rather than 
> > > > the metadataIdRef as it is for the other perfSONAR services). The 
> > > > same goes for the metadata id attribute.
> > > OK, we will try to support your suggestions.
> > >
> > > Is there a documentation about the requirements for a proper 
> > > perfSONAR service available, containing the minimal/optimum set of 
> > > interaction a service is supposed to support? This would make things 
> > > easier for us.
> > >
> > > Cheers,
> > >
> > > M&M
> > > > This is it for now. I'm available for any question or clarification.
> > > >
> > > >
> > > > Best Regards,
> > > >
> > > >
> > > > Michalis Michael