Flink安装部署

通用配置

• 解压

  • tar -zxf /chinaskills/flink-1.14.0-bin-scala_2.12.tgz -C /opt/
    

• 配置环境变量

  • vi /root/.bash_profile
    export FLINK_HOME=/opt/flink-1.14.0
    export PATH=$PATH:$FLINK_HOME/bin
    

local模式安装

• 启动

  • #切换到bin目录下
    ./start-cluster.sh
    
    #启动节点
    17917 TaskManagerRunner
    17648 StandaloneSessionClusterEntrypoint
    

• 提交

  • flink run ../examples/batch/WordCount.jar
    
    (whether,1)
    (whips,1)
    (who,2)
    (whose,1)
    (will,1)
    (wish,1)
    (with,3)
    (would,2)
    (wrong,1)
    (you,1)
    

Standalone模式安装

• 修改配置文件

  • 修改flink-conf.yaml文件

    • jobmanager.rpc.address: master
      

  • 修改masters文件

    • master:8081
      

  • 修改works文件

    • master
      slave1
      slave2
      

• 分发环境变量和软件目录

  • scp /root/.bash_profile slave1:/root/.bash_profile 
    scp /root/.bash_profile slave1:/root/.bash_profile
    scp -r flink-1.10.2 slave1:$PWD
    scp -r flink-1.10.2 slave2:$PWD
    

• 启动

  • #切换到bin目录下
    ./start-cluster.sh
    

• 提交运行（取结果的后十行）

  • flink run -m master:8081 -p 2 ../examples/batch/WordCount.jar
    (weary,1)
    (when,2)
    (whips,1)
    (who,2)
    (whose,1)
    (will,1)
    (wish,1)
    (would,2)
    (wrong,1)
    (you,1)
    

yarn模式安装

• 添加环境变量

  • vi /root/.bash_profile
    export HADOOP_CLASSPATH=$(/opt/hadoop-3.1.3/bin/hadoop classpath)
    

• on yarn 模式提交运行

  • flink run -m yarn-cluster -p 2 -yjm 2G -ytm 2G $FLINK_HOME/examples/batch/WordCount.jar
    (weary,1)
    (when,2)
    (whips,1)
    (who,2)
    (whose,1)
    (will,1)
    (wish,1)
    (would,2)
    (wrong,1)
    (you,1)