zoukankan      html  css  js  c++  java
  • Spark Standalone Mode Configuration

      For currently popular distributed framework Spark, here it shows the intro and steps to configure the spark standalone mode on several machines.

    It is easy to configure it from stratch.  The following instruction I note down is based on the spark-2.0.2-bin-hadoop2.7 as example on the linux debian machines for scala programming.

    Assume you have two machines with IP: 192.168.0.51 and 192.168.0.52 

    1.  Preinstall java, scala, sbt

      check: https://www.scala-lang.org/download/install.html        

                 http://www.scala-sbt.org/0.13/docs/Installing-sbt-on-Linux.html

    2. Download prebuilt spark version with hadoop. or you can compile on your own 

    the link can be referenced: https://spark.apache.org/downloads.html

    3. Unzip the file and create the link for easy visit later

    e.g.   execute: ln -s /usr/local/spark-2.0.2-bin-hadoop2.7 /usr/local/spark
     
    4. Configure the spark environments:
     (1) configure slaves file:   /usr/local/spark-2.0.2-bin-hadoop2.7/conf/slaves
    # A Spark Worker will be started on each of the machines listed below.
    192.168.0.51
    192.168.0.52
     
    (2) configure spar_env.sh.               e.g.
    #spark-env.sh
    export SCALA_HOME=/usr/local/scala
    export JAVA_HOME=/home/local/jdk
    #export SPARK_LOCAL_IP=localhost
    export SPARK_EXECUTOR_MEMORY=6g
    export SPARK_EXECUTOR_CORES=6
    export SPARK_MASTER_IP=192.168.0.51
    export SPARK_MASTER_PORT=8070
    export SPARK_MASTER_WEBUI_PORT=8080
    #export SPARK_WORKER_INSTANCES=1
    export SPARK_WORKER_PORT=8092
    #export SPARK_WORKER_MEMORY=4g
    #export SPARK_WORKER_CORES=4
     
    5.  Set up passwordless ssh access key
      (1) Generate ssh key without password
    $ ssh-keygen -t rsa -P ""
    

    (2) Copy id_rsa.pub to authorized-keys

    $  cat $HOME/.ssh/id_rsa.pub >> $HOME/.ssh/authorized_keys
    

    (3) Start ssh localhost           if you want to work in only one localhost machine for spark standalone

     $ ssh localhost
     
    6. Start spark
    $SPARK_HOME/sbin/start-all.sh
     execute  jps  to check worker and master have been up
     
    7. Write and run your application
     
    execute:  sbt package 
     
    execute:  $SPARK_HOME/bin/spark-submit
          --class "main.scala.MainAppTest"
          --master local[4]
          xxxxxxxx.jar 
     
     
     
     
     
  • 相关阅读:
    java foreach遍历的前提条件
    Java中的null值总结
    mybatis不可忽略的细节
    设计模式:创建型模式
    设计模式(四):原型模式
    设计模式(三):建造者模式
    设计模式(二):单例模式(DCL及解决办法)
    设计模式(一):简单工厂、工厂模式、抽象工厂
    定时任务 ScheduledExecutorService
    快速访问GitHub
  • 原文地址:https://www.cnblogs.com/anxin6699/p/7150048.html
Copyright © 2011-2022 走看看