Design of independent parallel sequences

A forum for discussing DataStage<sup>®</sup> basics. If you're not sure where your question goes, start here.

Moderators: chulett, rschirm, roy

Post Reply
franek
Participant
Posts: 4
Joined: Wed Jul 05, 2017 3:28 am
Location: Warsaw

Design of independent parallel sequences

Post by franek »

I have multiple sequences that looks like this (let's call it SimpleSequence):

WaitForFile -> JobActivity(parallel) -> ExecuteCommandActivity -> JobActivity(parallel)

What I want to achieve:
Those jobs concern one functional area and I want to schedule them once. To avoid scheduling each job I have to aggregate them in one main sequence job. But my other requirement is to make them independent in the context of errors. Is it ok to aggregate sequences in another sequence? I was thinking about main sequence job which will be composed of multiple SimpleSequences like this:


WaitForFile -> JobActivity1(parallel) -> ExecuteCommandActivity -> JobActivity2(parallel) (call it path 1)


WaitForFile -> JobActivity1(parallel) -> ExecuteCommandActivity -> JobActivity2(parallel) (call it path 2)

(...)

WaitForFile -> JobActivity1(parallel) -> ExecuteCommandActivity -> JobActivity2(parallel) (call it path n)

Problem is, that if I'm using checkpoints and one JobActivity1 aborts, best scenario which I can configure is to finish started activities and then whole sequence aborts. This causes, that if in some path JobActivity1 fails, propably any of JobActivity2 will finish. When checkpoints are disabled, all paths finish and main sequence finishes ok - I would like it finish not ok. Do you have any suggestions on my design? Is this even possible or I have to schedule them separatelly?
Post Reply