Hi, Renat!


*Suggestions*
> *1*. Define "input" and "output" at task level like this:
>      createVM:
>        input:
>          image_id: $.image_id
>        output: vm_id
> Where "output: vm_id" is basically a replacement for "store-as: vm_id" at
> action level, i.e. it's a hint to Mistral to store the output of this task
> under "vm_id" key in execution context. Again, the idea is to define task
> and action responsibilities more strictly:
>
>    - *Task is a high-level workflow building block which defines workflow
>    logical step and how it modifies workflow data. Task doesn't contain
>    technical details on how it's implemented.*
>
>
>    - *Action is an implementor of the workflow logical step defined by a
>    task. Action defines specific algorithm of how task is implemented.*
>
>
> *2*. User "parameters" only for actions to specify their additional
> properties influencing their nature (like method for HTTP actions).


Just clarify your thoughts:
 - All static keys and parameters should be in actions
 - All dynamic keys and parameters should be in tasks block
 - We may use our context to define some parameters in action or task

And, yes, I think it is a good idea to differentiate this. It is become
easier
_______________________________________________
OpenStack-dev mailing list
[email protected]
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Reply via email to