perfsonar-dev - Re: [pS-dev] Re: Functional Testing the E2EMon MP
Subject: perfsonar development work
List archive
- From: Michael Michalis <>
- To: Nicolas Simar <>
- Cc: Loukik Kudarimoti <>, Verena Venus <>, "Matthias K. Hamm" <>, Michael Michalis <>, Luis Marta <>, "Matthias.Hamm" <>, "Mark.Yampolskiy" <>, "" <>, Szymon Trocha <>
- Subject: Re: [pS-dev] Re: Functional Testing the E2EMon MP
- Date: Mon, 16 Jul 2007 10:30:12 +0300
Nicolas Simar wrote:
Some parts like echo requests are common between services so already made functional tests can be used. LS registration so far was tested manually Result codes are also kind of unique for each service. Doc/lit style can be checked indirectly when checking for the echo request ( a service not using doc/lit will return a soap fault). Logging is not currently directly checked and wsdl files produced at least for the java services are too general to check.
Loukik Kudarimoti wrote:
Verena Venus wrote:
Ok, we weren't aware of your new micro-release. Can you drop an email to release management whenever you produce a new micro-release?
Hi Loukik,
Am Monday 09 July 2007 13:16 schrieb Loukik Kudarimoti:
[...]
> 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.
>
In fact we do support this feature in the 0.2.1 release of our service.
> 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?
>
I'd like to bring up the idea of a perfSONAR validator (again), which should be easy to accomplish, now that we have the SOAP UI test projects for the services. Just extract what all these test files have in common, and you have testing script for common perfsonar services. This could also be the starting point for defining what a "proper" perfSONAR service is.
And obviously, we have a need for such a thing (see Mark and Matthias).
I take your point on the need for such a validtor. However, I think such a validator would still need a documented definition of what exactly does it need to validate (i.e., what features should it be checking for). I think we need an initial list (like echo functionality, ls registration, result codes, doc/lit style, logging, wsdl, etc)
That sounds like a reasonable starting point at those functionalities are used by multiple services.
Loukik and Michalis, have you got an idea on how we could proceed? Can we re-used some existing part of the functional tests?
If we are to have a perfSONAR service validator we must first put down a list of the features and characteristics a perfSONAR service should have. If the list is something like the above then i suggest that we develop a set of tests around the echo request along with tests verifying the LS registration(I'm not sure how to test logging).
So I propose a pre-testing stage where the service is checked whether it fits in perfSONAR minimum specifications and requirements and if so then proceed to further testing.
This set of tests can be build around the echo request. Using the echo request we can determine if the service actually uses doc/lit style, if it can respond with a result code and not a soap fault and of course the echo request functionality. We must also develop an automated way to see if a service has the ability to register to the LS(A query to a certain LS after deployment and after undeployment may do the trick). And we also need to see if further tests need to be developed for any other requirements.
#OFF TOPIC#
Its quite interesting because I'm currently investigating a tool brought under my attention by Andreas. Its actually a tool used for deploying web services with a build in testing mechanism which may came handy to us. http://smartfrog.org/presentations/distributed_testing_with_smartfrog_slides.pdf
#OFF TOPIC#
-Michalis
Cheers,
Nicolas
Loukik.
Kind regards,
Verena
--
Verena Venus - DFN Labor
Regionales Rechenzentrum Universitaet Erlangen-Nuernberg
Martensstrasse 1, D-91058 Erlangen
- Re: Functional Testing the E2EMon MP, Loukik Kudarimoti, 07/09/2007
- Re: [pS-dev] Re: Functional Testing the E2EMon MP, Verena Venus, 07/09/2007
- Re: [pS-dev] Re: Functional Testing the E2EMon MP, Loukik Kudarimoti, 07/09/2007
- Re: [pS-dev] Re: Functional Testing the E2EMon MP, Verena Venus, 07/09/2007
- Re: [pS-dev] Re: Functional Testing the E2EMon MP, Nicolas Simar, 07/16/2007
- Re: [pS-dev] Re: Functional Testing the E2EMon MP, Michael Michalis, 07/16/2007
- Topology Tests, murilo, 07/16/2007
- Re: [pS-dev] Re: Functional Testing the E2EMon MP, Szymon Trocha, 07/18/2007
- Re: [pS-dev] Re: Functional Testing the E2EMon MP, Loukik Kudarimoti, 07/09/2007
- Re: Functional Testing the E2EMon MP, Szymon Trocha, 07/10/2007
- Re: Functional Testing the E2EMon MP, Matthias K. Hamm, 07/10/2007
- Re: Functional Testing the E2EMon MP, Loukik Kudarimoti, 07/10/2007
- Re: [pS-dev] Re: Functional Testing the E2EMon MP, Verena Venus, 07/09/2007
Archive powered by MHonArc 2.6.16.