perfsonar development work

[Verena Venus <>]

Hi *,

It seems that we are converging :)

So I like to finalize this, which means, putting the alternate approach 
(according to the examples from Jason and me) in a state of first defined 
behaviour for requesting (and responding to) self-tests.

Thus, I'm going to edit the wiki site, providing the BWCTL example with full 
request and response (and correct unique ids), and write a bit about the 
general behaviour. This should then be incorporated in the specification as 
extension to EchoRequest.

So, if anyone has serious objections speak now or remain silent forever ;)

Note that this is only to start with the proper behaviour of selftesting in 
generell for a generic request and with individual error/success report for 
each selftest.

Everything else, including the dicussion about usefulness of certain types of 
tests is not subject of this specification!

Regards,
Verena
 
Am Dienstag, 24. Juni 2008 11:39 schrieb Jason Zurawski:
> All;
>
> >> sends back conflicting messages as to the results of the test (why
> >> are there 2 errors and a warning, it appears that again we have run 3
> >> separate tests here...).  It also is imposing additional result
> >> semantics to the datum items, if this was an 'error' or a 'waring'
> >> the eventType should say so.  If this proposal is being offered to
> >> extend syslog information mining it is being done in the wrong
> >> manner, there would need to be 3 sub-results (note the altered
> >> eventTypes and single datum elements) as per the original example
> >> where the selftest itself was code for several tests:
> >
> > I'm also noticing that the datum element can't be used that way:
> >
> > <nmwgr:datum *value="warning"*>*Could not read configuration falling
> > back to defaults*</nmwgr:datum>
> >
> > the value attribute is being set two times with different meanings.
> > The schema allows any attribute on datum (i.e. change value to type),
> > but as Jason is pointing, the type of information should be defined by
> > the eventType.
>
> Thanks for catching that, you are absolutely right (I should have
> proofread my example better):
>
>  <nmwg:message id="message1213971959" type="EchoResponse"
> xmlns:nmwg="http://ggf.org/ns/nmwg/base/2.0/";
> xmlns:result="http://ggf.org/ns/nmwg/result/2.0/";>
>    <nmwg:metadata id="meta">
>      <nmwg:eventType>
> http://schemas.perfsonar.net/tools/admin/selftest/MP/BWCTL/well_configured/
>WARNING/1.0 </nmwg:eventType>
>    </nmwg:metadata>
>    <nmwg:data id="causes" metadataIdRef="meta">
>      <nmwgr:datum>Could not read configuration falling back to
> defaults</nmwgr:datum>
>    </nmwg:data>
>  </nmwg:message>
>
>  <nmwg:message id="message1213971959" type="EchoResponse"
> xmlns:nmwg="http://ggf.org/ns/nmwg/base/2.0/";
> xmlns:result="http://ggf.org/ns/nmwg/result/2.0/";>
>    <nmwg:metadata id="meta">
>       <nmwg:eventType>
> http://schemas.perfsonar.net/tools/admin/selftest/MP/BWCTL/well_configured/
>ERROR/1.0 </nmwg:eventType>
>    </nmwg:metadata>
>    <nmwg:data id="causes" metadataIdRef="meta">
>      <nmwgr:datum>bwctld executable not found!</nmwgr:datum>
>    </nmwg:data>
>  </nmwg:message>
>
>  <nmwg:message id="message1213971959" type="EchoResponse"
> xmlns:nmwg="http://ggf.org/ns/nmwg/base/2.0/";
> xmlns:result="http://ggf.org/ns/nmwg/result/2.0/";>
>    <nmwg:metadata id="meta">
>      <nmwg:eventType>
> http://schemas.perfsonar.net/tools/admin/selftest/MP/BWCTL/well_configured/
>ERROR/1.0 </nmwg:eventType>
>    </nmwg:metadata>
>    <nmwg:data id="causes" metadataIdRef="meta">
>      <nmwgr:datum>bwctld not running!</nmwgr:datum>
>    </nmwg:data>
>  </nmwg:message>
>
> -jason