The Hortonworks Community Connection is now live. A completely rebuilt Q&A forum, Knowledge Base, Code Hub and more, backed by the experts in the industry.

You will be redirected here in 10 seconds. If your are not redirected, click here to visit the new site.

The legacy Hortonworks Forum is now closed. You can view a read-only version of the former site by clicking here. The site will be taken offline on January 31,2016

YARN Forum

Allow mixed machine hardware specification

  • #46980
    Feng Meng
    Participant

    Hi, does Yarn allow different machine hardware specification e.g. one node with 12 core and 256G memory and later add another node with 16 core and 512G memory?

  • Author
    Replies
  • #46982
    Rohit Bakhshi
    Moderator

    Hi Feng,

    Yes, YARN allows you to incorporate mixed hardware specifications for the nodes. You can set a different configuration in yarn-site.xml for each node.

    For example, for the node with 256 GB RAM, you can tell the Node Manager to use up to 220 GB for scheduling containers, and for the node with 52 GB RAM, you can tell the Node Manager to use up to 440 GB for scheduling containers.

    The Resource Manager will take this all into account when it allocates resources across the Node Managers.

    #46988
    Feng Meng
    Participant

    Thanks, good to know. Does this mean I need to modify yarn-site.xml for every node based on its specific hardware spec? Originally, I thought yarn-site.xml is shared by all nodes. For example, I should set following parameters differently for each node, correct?

    <property>
    <name>yarn.nodemanager.resource.memory-mb</name>
    <value>32768</value>
    </property>
    <property>
    <name>yarn.nodemanager.resource.cpu-vcores</name>
    <value>16</value>
    </property>

    #47617

    That is correct.

The forum ‘YARN’ is closed to new topics and replies.

Support from the Experts

A HDP Support Subscription connects you experts with deep experience running Apache Hadoop in production, at-scale on the most demanding workloads.

Enterprise Support »

Become HDP Certified

Real world training designed by the core architects of Hadoop. Scenario-based training courses are available in-classroom or online from anywhere in the world

Training »

Hortonworks Data Platform
The Hortonworks Data Platform is a 100% open source distribution of Apache Hadoop that is truly enterprise grade having been built, tested and hardened with enterprise rigor.
Get started with Sandbox
Hortonworks Sandbox is a self-contained virtual machine with Apache Hadoop pre-configured alongside a set of hands-on, step-by-step Hadoop tutorials.
Modern Data Architecture
Tackle the challenges of big data. Hadoop integrates with existing EDW, RDBMS and MPP systems to deliver lower cost, higher capacity infrastructure.