First and foremost, I am a Control-M consultant. Control-M is a job scheduler that has integrations with many different applications (SAP, Peoplesoft etc.) and the ability to define your own integrations to other applications. I have a customer that is using MS Dynamix AX, but would like to integrate their workload from Dynamix with Control-M.
I am trying to determine feasibility, and if so what is required or what is the best way.
From my understanding, job scheduling is defined at the batch job level in AX. Job dependencies, alerts, and actions are defined in a batch task as well as the actual command or code to execute. As far as terminology is concerned, a job in AX is like a folder (container) in Control-M, and a task in AX is a job in Control-M.
With this in mind, I have two possible scenarios. In both cases -> in AX there would be NO scheduling defined in the batch job.
1. The batch job would be launched using some AX API, Webservice, cli or what is the preferred, simplest, most effective way. Batch tasks would be defined in AX as part of these jobs, and the job status in Control-M would depend on the batch job status. We can performing corrective actions on a job (kill, rerun) in Control-M, and can send custom actions to AX if this functionality is possible.
OR
My preferred way:
2. A job in Control-M would correspond to a batch task (since within Control-M we can implement dependencies , schedule and all that fun stuff). I do not know if batch tasks can be defined independent of batch job, or if they are defined as a different entity altogether and linked to a batch task. The idea would be to invoke this task (api, webservice, cli and so on), and treat it's execution as an individual job in Control-M. I prefer this way, since I believe it makes taking corrective action on an individual task much simpler, as opposed to having an operator have to engage in AX within a job construct. I do not 'know' for sure if that is the best approach in reality.
So AX experts, what do you think?