Post by fish on Oct 19, 2018 9:04:44 GMT
Some questions related to Tasks and Task payload.
1. Description of property StartTime talks about "The date-time stamp that the task was last started."Why is word "last" used here? I assume that a new separate Task resource is created each time a specific asynchronous operation is executed. So how could one Task resource instance be started more than once? Or is this only a problem with wording? (Same for EndTime)
2. Wording in sub properties of Task Payload suggest also that a specific Task can be restarted. Example: HttpOperation: "The HTTP operation to perform to execute this Task." This gets confusing in similar way I think. I would have preferred something like "The HTTP operation that initiated this Task." I assume that Payload object should expose which asynchronous operation that originally caused this Task resource to be created and exposed. (Which should be a one-time thing.) Could you please confirm that this is the intended usage of Payload object in Task resource?
Questions 1 and 2 comes down to the following question I think: Is a new Task created each time a specific asynchronous operation is executed? Or could the same the Task resource (same id) be "restarted" if the same asynchronous operation is executed again? I always assumed that there is always a new Task resource created (and new Task Monitor) for each asynchronous operation execution. Otherwise the execution outcome of the first operation would be lost if I execute the same asynchronous operation again... (Or is this actually implementation specific, and to be selected by the service provider/vendor?) If my assumption is still correct I think you should consider to rephrase these texts in the Task schema.
3. LongDescription of new property HidePayload ends with "If this property is not specified when the Task is created, the default value shall be False." What does this mean? I assume the end-users /clients cannot create Tasks explicitly, as Task resources are created automatically by the system/service when an asynchronous operation is initiated. So how can this default value be an issue if the Task can never be created (e.g. via POST on Task collection by API client)? I assume that the service provider (vendor) should select to expose the HidePayload value that is most suitable for each type of asynchronous operation?
4. Task Description in schema says "This resource contains information about a specific Task scheduled by or being executed by a Redfish service's Task Service.". Why is the word "scheduled" used? This could also indicate that Tasks resources are possible to be explicitly created by end-user /client, and then also scheduled to start at a specific time. But I assume that this is not the case. OK? Should not Jobs be used for this purpose?
1. Description of property StartTime talks about "The date-time stamp that the task was last started."Why is word "last" used here? I assume that a new separate Task resource is created each time a specific asynchronous operation is executed. So how could one Task resource instance be started more than once? Or is this only a problem with wording? (Same for EndTime)
2. Wording in sub properties of Task Payload suggest also that a specific Task can be restarted. Example: HttpOperation: "The HTTP operation to perform to execute this Task." This gets confusing in similar way I think. I would have preferred something like "The HTTP operation that initiated this Task." I assume that Payload object should expose which asynchronous operation that originally caused this Task resource to be created and exposed. (Which should be a one-time thing.) Could you please confirm that this is the intended usage of Payload object in Task resource?
Questions 1 and 2 comes down to the following question I think: Is a new Task created each time a specific asynchronous operation is executed? Or could the same the Task resource (same id) be "restarted" if the same asynchronous operation is executed again? I always assumed that there is always a new Task resource created (and new Task Monitor) for each asynchronous operation execution. Otherwise the execution outcome of the first operation would be lost if I execute the same asynchronous operation again... (Or is this actually implementation specific, and to be selected by the service provider/vendor?) If my assumption is still correct I think you should consider to rephrase these texts in the Task schema.
3. LongDescription of new property HidePayload ends with "If this property is not specified when the Task is created, the default value shall be False." What does this mean? I assume the end-users /clients cannot create Tasks explicitly, as Task resources are created automatically by the system/service when an asynchronous operation is initiated. So how can this default value be an issue if the Task can never be created (e.g. via POST on Task collection by API client)? I assume that the service provider (vendor) should select to expose the HidePayload value that is most suitable for each type of asynchronous operation?
4. Task Description in schema says "This resource contains information about a specific Task scheduled by or being executed by a Redfish service's Task Service.". Why is the word "scheduled" used? This could also indicate that Tasks resources are possible to be explicitly created by end-user /client, and then also scheduled to start at a specific time. But I assume that this is not the case. OK? Should not Jobs be used for this purpose?