Apache Oozie

Chapter 7. Data Trigger Coordinator

In Chapter 6, we primarily discussed how Oozie materialized coordinator actions at periodic intervals and subsequently executed the workflow. In other words, we only considered the time-based trigger to start workflows. However, time is not the only dependency that determines when to launch a workflow for many use cases. A common use case is to wait for input data. If a workflow is started before its required data is available, the workflow execution will either produce wrong results or fail. The Oozie coordinator allows users to express both data and time dependency together and to kick off the workflow accordingly. There are many diverse use cases based on data dependency that poses serious challenges to the design of the coordinator. In this chapter, we explain how to express both data and time dependency in a coordinator and how Oozie manages the workflow executions.

Expressing Data Dependency

It’s important to understand the three terms, dataset, input-events, and output-events, that Oozie uses to describe data dependencies in a coordinator XML.

Dataset

A dataset is a logical entity to represent a set of data produced by an application. A user can define a dataset either using its directory location or using metadata. Oozie has always supported directory-based data dependency. Recently, Oozie introduced metadata-based data dependency as well. This book primarily focuses on the directory-based dataset, as that’s the most commonly used approach. ...