Ambari Forum

Cannot start namenode in Ambari

  • #48181
    Jason Williams
    Participant

    I am unable to start the namenode on my cluster using Ambari. I am getting several errors, which include.

    • notice: /Stage[2]/Hdp-hadoop::Initialize/Configgenerator::Configfile[hdfs-site]/File[/etc/hadoop/conf/hdfs-site.xml]/content: content changed ‘{md5}….0′ to ‘{md5}….5′
    • notice: /Stage[2]/Hdp-hadoop::Initialize/Configgenerator::Configfile[capacity-scheduler]/File[/etc/hadoop/conf/capacity-scheduler.xml]/content: content changed ‘{md5}…..d’ to ‘{md5}…..3′
    • notice: /Stage[2]/Hdp-hadoop::Initialize/Configgenerator::Configfile[capacity-scheduler]/File[/etc/hadoop/conf/capacity-scheduler.xml]/owner: owner changed ‘yarn’ to ‘hdfs’
    • notice: /Stage[2]/Hdp-hadoop::Initialize/Configgenerator::Configfile[core-site]/File[/etc/hadoop/conf/core-s
    • ERROR: Namenode directory(s) is non empty. Will not format the namenode. List of non-empty namenode dirs
    • returned 1 instead of one of [0] at /var/lib/ambari-agent/puppet/modules/hdp-hadoop/manifests/namenode/format.pp:53

    • /Anchor[hdp::exec::set namenode mark::begin]: Dependency Exec[/tmp/checkForFormat.sh] has failures: true
    • /Exec[set namenode mark]: Dependency Exec[/tmp/checkForFormat.sh] has failures: true
    • Exec[mkdir -p /var/run/hadoop/hdfs]/Anchor[hdp::exec::mkdir -p /var/run/hadoop/hdfs::begin]: Dependency Exec[/tmp/checkForFormat.sh] has failures: true
    • Exec[mkdir -p /var/log/hadoop/hdfs]/Exec[mkdir -p /var/log/hadoop/hdfs]: Dependency Exec[/tmp/checkForFormat.sh] has failures: true
    • Exec[mkdir -p /var/log/hadoop/hdfs]/Anchor[hdp::exec::mkdir -p /var/log/hadoop/hdfs::end]: Dependency Exec[/tmp/checkForFormat.sh] has failures: true

    If this is not the correct forum to post this, please let me know. Thank you

to create new topics or reply. | New User Registration

  • Author
    Replies
  • #48226
    Kenny Zhang
    Moderator

    Hi Jason,

    Is this the first time you bring up Namenode from Ambari? If yes, please let me know whether this is a fresh and clean install or you’re performing a Ambari takeover.

    I suspect that your dfs.name.dir is not clean, which is specify in the hdfs-site.xml should be empty in the first time for the NN format. If this should a fresh install, then you might want to make sure the dfs.name.dir is empty first.

    Please note dfs.name.dir is deprecated by dfs.namenode.name.dir in HDP 2.

    Thanks,
    Kenny

    #48285
    Jason Williams
    Participant

    No, this is not the first time we are starting the namenode from ambari. right now the only folder inside of the namenode directory is current.

    Can i safely delete the “current” folder and it’s contents?

    Thank you,
    Jason

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.