[ogsa-hpcp-wg] Platform Terminate Semantics

[Christopher Smith]

I also noticed that I could "cancel" already finished jobs. I just decided
in my implementation that it seems strange to cancel an already finished
job, and maybe a bit misleading to a client (perhaps something isn't
properly cleaned if you cancel compared to if you finish).

Anyway ... my chosen semantics ... the spec says nothing on this point, and
perhaps should.

-- Chris


On 09/11/06 10:34, "Peter G. Lane" <lane at mcs.anl.gov> wrote:

> When I try to destroy a Finished job through the Platform endpoint, it always
> returns the following
> error:
> 
> java.lang.Exception: Unable to destroy job:
> Activity identified by EPR has finished already
> 
> Is this because the service doesn't have any per-job state? Are these
> semantics acceptable. The BES
> spec doesn't really say anything about the semantics of calling this operation
> after the job status
> is Finished. I treats it more like a cancel operation. The description of the
> Finished state also
> doesn't say anything about what this means with respect to a job's state.
> 
> Anyway, I'm not trying to pick on Platform for doing it this way, but I think
> it's an interesting
> question that should be addressed at some point.
> 
> Peter
> --
>   ogsa-hpcp-wg mailing list
>   ogsa-hpcp-wg at ogf.org
>   http://www.ogf.org/mailman/listinfo/ogsa-hpcp-wg