Book Image

Learning YARN

Book Image

Learning YARN

Overview of this book

Today enterprises generate huge volumes of data. In order to provide effective services and to make smarter and more intelligent decisions from these huge volumes of data, enterprises use big-data analytics. In recent years, Hadoop has been used for massive data storage and efficient distributed processing of data. The Yet Another Resource Negotiator (YARN) framework solves the design problems related to resource management faced by the Hadoop 1.x framework by providing a more scalable, efficient, flexible, and highly available resource management framework for distributed data processing. This book starts with an overview of the YARN features and explains how YARN provides a business solution for growing big data needs. You will learn to provision and manage single, as well as multi-node, Hadoop-YARN clusters in the easiest way. You will walk through the YARN administration, life cycle management, application execution, REST APIs, schedulers, security framework and so on. You will gain insights about the YARN components and features such as ResourceManager, NodeManager, ApplicationMaster, Container, Timeline Server, High Availability, Resource Localisation and so on. The book explains Hadoop-YARN commands and the configurations of components and explores topics such as High Availability, Resource Localization and Log aggregation. You will then be ready to develop your own ApplicationMaster and execute it over a Hadoop-YARN cluster. Towards the end of the book, you will learn about the security architecture and integration of YARN with big data technologies like Spark and Storm. This book promises conceptual as well as practical knowledge of resource management using YARN.
Table of Contents (20 chapters)
Learning YARN
Credits
About the Authors
Acknowledgments
About the Reviewers
www.PacktPub.com
Preface
Index

High-level changes from MRv1 to MRv2


With the introduction of YARN, the architecture for Hadoop job execution and management framework changed. In this section, we'll discuss the list of high-level changes observed in MRv2 framework.

The evolution of the MRApplicationMaster service

In YARN, the responsibility of JobTracker is divided across the ResourceManager service and application-specific ApplicationMaster service. For management of MapReduce jobs, MRApplicationMaster service is defined in the Hadoop framework. For each MapReduce job submitted to ResourceManager, an instance MRApplicationMaster service is launched. After successful execution of the job, the MRApplicationMaster service is terminated.

The MRApplicationMaster service is responsible for:

  • Registering the job with the ResourceManager

  • Negotiating YARN containers for execution of map reduce tasks

  • Interacting with NodeManager to manage execution of allocated containers

  • Handling task failure and reinitiate failed tasks

  • Handling client...