HDP on Windows – Other Forum

Map Reduce Job Performance on a multi node cluster

  • #50538
    SambaSivaRao Y
    Participant

    Hi,

    We are facing performance issue while running a map reduce job on 4 node cluster compared to Single node.
    The job on multi node cluster is taking more time compared to Single node.

    Details of job’s run time :
    On a Single node : 26 min
    On a 4 node cluster : >35 min.

    Details of Single node :
    8 GB RAM

    Details of Multi node cluster :
    Master node : 8 GB RAM
    2 Data Nodes with 8GB RAM & 2 with 4 GB.

    We are also struggling to monitor the running job on multi node cluster.
    Please guide me on this.

    Thanks,
    SambaShiva.

to create new topics or reply. | New User Registration

  • Author
    Replies
  • #52098
    SambaSivaRao Y
    Participant

    Issue resolved, We are running the job from Eclipse instead of Command line.
    This is answer for another question at
    http://hortonworks.com/community/forums/topic/map-reduce-job-run-from-eclipse/

    #53727
    IngEnsi
    Participant

    Hi SambaSivaRaoy ,
    I’am running a mapreduce job on 3 node cluster and im facing the same problem.
    The job on multinode cluster is taking more time compared to single node (35 min vs 51 min) , I noticed that the first 3 mappers run within seconds but the last one is taking minutes. Is it a problem of configuration??

    Details:
    Master Node 3GB RAM
    Slave1 Node 3GB RAM
    Slave2 Node 4GB RAM

    #56163
    SambaSivaRao Y
    Participant

    Hi,

    Could you please provide me the details of job type (Whether you are loading data from DB or to DB ).

    Thanks,
    SambaSiva.

You must be to reply to this topic. | Create Account

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.