Apache Hadoop YARN

by | Sep 12, 2018 | Apache Hadoop, Big Data | 0 comments

Yarn definition

Yarn (Yet Another Resource negotiator) is a data operating system and distributed Resource Manager, also known as Hadoop 2 as it is the evolution of Hadoop Map-Reduce.

The most significant changes of Hadoop 2 over Hadoop 1 is that the thread technology is included, this technology provides an effective allocation of resources, for it runs two demons:

Resource Manager: responsible for managing the resources on their multiple nodes.

Application Master: responsible for negotiating resources with the resource manager on the one hand and with the node manager to run and monitor for another. There will be a master application for each “job” sent to the cluster.

Architecture

Hadoop YARN achitecture

Hadoop YARN achitecture

View Yarn Processes

The service of Yarn in cloudy or Hortoworks is usually configured in Port 8088

http://localhost:8088 or http://127.0.0.1:8088

 

Commands

Commands used in the Hortonworks shell:
Restore Service:

sudo service Ambari-server restart

View the Process list

yarn application -list -appStates ALL

Note You can specify the processes by the state by changing ALL by: NEW, NEW_SAVING, submitted, accepted, RUNNING, finished, failed, killed.

Kill applications ripped off

yarn application -kill

View Logs Concrete Application

yarn logs -applicationId <id-aplicación>

 

Source: Official website

Source: Commands

0 Comments

Submit a Comment

Your email address will not be published. Required fields are marked *