HDFS Forum

NameNode functions

  • #34199
    Swapnil Patil
    Participant

    Does namenode writes anything on datanode? I am confused in the following things..
    1) If I have set up replication factor to 3..then does namenode writes these replicas to 3 datanodes? or 1 datanode writes to other datanodes?

    Thanks inadvance

to create new topics or reply. | New User Registration

  • Author
    Replies
  • #34235
    Ravi Kiran
    Member

    Hi Swapnil,
    The namenode never , ever writes blocks to a datanode directly. The data nodes form a pipeline and the first one sends the block to the second and so on downstream.
    You can see the process of a FileWrite from here
    http://books.google.co.in/books?id=drbI_aro20oC&pg=PA70&lpg=PA70&dq=anatomy+of+file+write+in+hadoop&source=bl&ots=tZDmtaj0d8&sig=Sl6aT4XVhHuXYxN_De7teMQoEu8&hl=en&sa=X&ei=RUMnUoSvE4_yiQKiy4C4Cg&ved=0CDwQ6AEwAw#v=onepage&q=anatomy%20of%20file%20write%20in%20hadoop&f=false

    Regards
    Ravi Magham

    #34341
    Swapnil Patil
    Participant

    Hi Ravi,
    Thanks for your reply..
    How one datanode recognizes other datanodes in the cluster? I mean which file they refer to for other datanodes IP addresses?
    Do they refer etc/hosts file or slaves file
    What kind of setings I need to do in the respective datanode config files.

    #39179
    Robert Molina
    Moderator

    Hi Swapnil,
    Have you tried just installing the cluster with Ambari? It makes it easier to set up the cluster and administer.

    Regards,
    Robert

The topic ‘NameNode functions’ is closed to new 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.