ambari meetup: yarn

17
© Hortonworks Inc. 2013 YARN and Ambari YARN service management using Ambari Srimanth Gunturi September 25 th , 2013 Page 1

Upload: hortonworks

Post on 26-Jan-2015

115 views

Category:

Technology


4 download

DESCRIPTION

YARN service management using Apache Ambari.

TRANSCRIPT

Page 1: Ambari Meetup: YARN

© Hortonworks Inc. 2013

YARN and AmbariYARN service management using Ambari

Srimanth Gunturi

September 25th, 2013

Page 1

Page 2: Ambari Meetup: YARN

© Hortonworks Inc. 2013

Agenda

• YARN Overview• Installing• Monitoring• Configuration• Capacity Scheduler• MapReduce2• Future

Page 2Architecting the Future of Big Data

Page 3: Ambari Meetup: YARN

© Hortonworks Inc. 2013

YARN Overview

Page 3Architecting the Future of Big Data

• Yet Another Resource Negotiator (YARN)• Purpose - Negotiating resources (Memory, CPU, Disk, etc.) on a cluster• What was wrong with original MapReduce?

– Cluster not open to non-MapReduce paradigms– Inefficient resource utilization (Map slots, Reduce slots)– Upgrade rigidity

• MapReduce (Hadoop 1.0) -> YARN + MapReduce 2 (Hadoop 2.0)

Page 4: Ambari Meetup: YARN

© Hortonworks Inc. 2013

YARN Overview - Applications

Page 4Architecting the Future of Big Data

• MapReduce1 applications are fully compatible with MapReduce2 – Same JARs can be used– Binary compatibility with org.apache.hadoop.mapred API– Source compatibility with org.apache.hadoop.mapreduce API

Page 5: Ambari Meetup: YARN

© Hortonworks Inc. 2013

YARN Overview - Architecture

Page 5Architecting the Future of Big Data

• ResourceManager• NodeManagers• Containers• Applications (ApplicationMasters)

Page 6: Ambari Meetup: YARN

© Hortonworks Inc. 2013

Installing

Page 6Architecting the Future of Big Data

Page 7: Ambari Meetup: YARN

© Hortonworks Inc. 2013

Monitoring

Page 7Architecting the Future of Big Data

Page 8: Ambari Meetup: YARN

© Hortonworks Inc. 2013

Monitoring – NodeManager Summary

Page 8Architecting the Future of Big Data

NodeManagers Status• Active – In communication with RM• Lost – Not communicating with RM• Unhealthy – Flagged by custom health

check script identified by propertyyarn.nodemanager.health-checker.script.path

• Rebooted – Automatically restarted due to internal problems

• Decommissioned – RM ignoring communications from host. Host placedin yarn.resourcemanager.nodes.exclude-path file.

Page 9: Ambari Meetup: YARN

© Hortonworks Inc. 2013

Monitoring – Container Summary

Page 9Architecting the Future of Big Data

Containers• Allocated – Containers which have

been created with requested resources.• Pending – Containers, whose resources

will become available and are pending creation.

• Reserved – Containers, whose resources are not yet available.

Examples10 GB Cluster• Request three 5GB containers

• 2 allocated, 1 pending.• Request three 4GB containers

• 2 allocated, 1 reserved (2GB)

Page 10: Ambari Meetup: YARN

© Hortonworks Inc. 2013

Monitoring – Applications Summary

Page 10Architecting the Future of Big Data

Applications• Submitted – Application requests made

to YARN.• Running – Application with Masters

which have been created and are running.• Pending – Application requests which

are pending creation.• Completed – Applications which have

completed running. They could havebeen successful, killed or failed.

• Killed – Applications which have beenterminated by user

• Failed – Applications which have failedto run due to internal failures.

Page 11: Ambari Meetup: YARN

© Hortonworks Inc. 2013

Monitoring – Memory Summary

Page 11Architecting the Future of Big Data

Cluster Memory• Used – Memory resource currently

being used across the cluster• Reserved – Memory resources that

are set aside for being allocated.• Total – Memory resource available

across entire cluster

Page 12: Ambari Meetup: YARN

© Hortonworks Inc. 2013

Monitoring – Alerts

Page 12Architecting the Future of Big Data

Service Alerts- ResourceManager health- % NodeManagers alive

Host Alerts- NodeManager health- NodeManager process check

Page 13: Ambari Meetup: YARN

© Hortonworks Inc. 2013

Monitoring – Graphs

Page 13Architecting the Future of Big Data

Page 14: Ambari Meetup: YARN

© Hortonworks Inc. 2013

Configuration

Page 14Architecting the Future of Big Data

Page 15: Ambari Meetup: YARN

© Hortonworks Inc. 2013

Configuration – Capacity Scheduler

Page 15Architecting the Future of Big Data

QueuesRooto Ao Bo C

o C1o C2

o default

• Hierarchical Queues• Capacity Guarantees

• Capacity (%)• Maximum-am-resource-percent (%)

• Elasticity• Maximum-capacity (%)

• Access Control

Page 16: Ambari Meetup: YARN

© Hortonworks Inc. 2013

MapReduce2

Page 16Architecting the Future of Big Data

YARN-321: Generic application history service

Page 17: Ambari Meetup: YARN

© Hortonworks Inc. 2013

Future

Page 17Architecting the Future of Big Data

• Support more YARN applications• Improve per application-type information• Improve Capacity Scheduler configuration• Better health checks