基于Yarn的Spark环境，统计哈姆雷特词频(1)

一、最流行的大数据框架Spark

1. Yarn 环境搭建
2. Spark History Server 以及 Yarn MapReduce History Servcer
3. Spark-submit 提交到Yarn 运行

二、Docker部署Hadoop Yarn

部署结果：

• 一台namenode节点，运行

namenode
resourcemanager
JobHistoryServer
HistoryServer

• 2台datanode节点，运行

datanode
nodemanager

• 主机Mac

docker宿主机（virtualbox）
Intellij idea
spark client
hdfs client

三、网络结构

• mac 192.168.99.1
• namenode 172.18.0.11
• datanode1,datanode2 172.18.0.13 172.18.0.14
• virtualbox 网桥 192.168.99.100

建立192.168.99.1 ~ 172.18.0.0 路由

sudo route -n add 172.18.0.0/24 192.168.99.100 

docker 创建 172.18 网段，命名hadoopnet，docker设置ip必须先创建网络

docker network create --subnet=172.18.0.0/16 hadoopnet 

四、启动docker

本地文件夹，整理好的工作空间

说明：每个文件夹包含一个启动脚本和一个hdfs挂载的共享卷data

说明：本地的hadoop目录会挂载到docker中hadoop/etc/hadoop配置文件目录。

1. NameNode

docker run --name namenode \ --hostname namenode \ --network hadoopnet \ --ip 172.18.0.11 \ -d \ -v $PWD/data:/opt/tmp \ -v /Users/wangsen/hadoop/datanode/hadoop:/opt/hadoop-2.7.3/etc/hadoop \ -v $PWD/spark-2.1.1-bin-hadoop2.7:/opt/spark \ --rm dbp/hadoop 

dbp/hadoop是docker镜像的名字，共加载了3个共享卷（文件夹）

• /opt/tmp hdfs 存储路径
• etc/hadoop hadpoop配置路径
• 主节点挂载spark

在创建镜像的时候没有装载spark，hadoop是通过Dockerfile创建dbp/hadoop时，装载到镜像中的；设置spark采用装载模式，也可以重新commit或build dockerfile生成包含spark的镜像。

2. DataNode(datanode1、datanode2)

docker run --name datanode1 --hostname datanode1 --network hadoopnet --ip 172.18.0.13 -d -v $PWD/data:/opt/tmp -v /Users/wangsen/hadoop/datanode/hadoop:/opt/hadoop-2.7.3/etc/hadoop --rm dbp/hadoop 

docker run --name datanode2 --hostname datanode2 --network hadoopnet --ip 172.18.0.14 -d -v $PWD/data:/opt/tmp -v /Users/wangsen/hadoop/datanode/hadoop:/opt/hadoop-2.7.3/etc/hadoop --rm dbp/hadoop 

五、启动HDFS、YARN

• etc/hadoop/core-site.xml

 ## 配置HDFS路径 <property> <name>fs.defaultFS</name> <value>hdfs://namenode:9000</value> </property> 

• etc/hadoop/hdfs-site.xml

 <property> <name>dfs.replication</name> <value>3</value> </property> <property> <name>dfs.namenode.name.dir</name> <value>/opt/tmp</value> </property> <property> <name>dfs.datanode.data.dir</name> <value>/opt/tmp</value> </property> 

• etc/hadoop/yarn-site.xml

 <property> <name>yarn.nodemanager.aux-services</name> <value>mapreduce_shuffle</value> </property> <property> <name>yarn.resourcemanager.address</name> <value>namenode:18040</value> </property> <property> <name>yarn.resourcemanager.scheduler.address</name> <value>namenode:18030</value> </property> <property> <name>yarn.resourcemanager.resource-tracker.address</name> <value>namenode:18025</value> </property> <property> <name>yarn.resourcemanager.admin.address</name> <value>namenode:18141</value> </property> <property> <name>yarn.resourcemanager.webapp.address</name> <value>namenode:18088</value> </property> <property> <name>yarn.log-aggregation-enable</name> <value>true</value> </property> <property> <name>yarn.log.server.url</name> <value>http://namenode:19888/jobhistory/logs</value> </property> <property> <name>yarn.nodemanager.vmem-check-enabled</name> <value>false</value> </property> <property> <name>yarn.nodemanager.pmem-check-enabled</name> <value>false</value> </property> 

• spark/conf/spark-env

export HADOOP_CONF_DIR=/opt/hadoop-2.7.3/etc/hadoop 

• spark/conf/spark-defaults.conf

## 配置spark ui 页面，通过yarn history服务查看spark任务运行结果 ## hdfs:///tmp/spark/events是hdfs上的路径，保存spark运行信息 spark.master=local spark.yarn.historyServer.address=namenode:18080 spark.history.ui.port=18080 spark.eventLog.enabled=true spark.eventLog.dir=hdfs:///tmp/spark/events spark.history.fs.logDirectory=hdfs:///tmp/spark/events 

• hadoop/etc/hadoop-env.sh

修改JAVA_HOME，填写java_home的绝对路径

启动顺序

• HDFS
  namenode -->sbin/hadoop-daemon.sh start namenode
  datanode -->sbin/hadoop-daemon.sh start datanode
  (已经设置好ssh免密码登录，docker共享了public_key文件。)
• Yarn
  namenode --> sbin/yarn-daemon.sh start resourcemanager
  datanode -->sbin/yarn-daemon.sh start nodemanager
• Spark jobserver
  namenode--> sbin/mr-jobhistory-daemon.sh start historyserver
  namenode--> spart/sbin/start-history-server.sh

六、浏览spark histroy页面

http://namenode:18080

附录 Dockerfile

如果你希望按作者的思路，搭建自己的spark docker集群，那么你可以从Dockerfile 创建image开始。

FROM ubuntu:16.04 MAINTAINER wsn RUN apt-get update RUN apt-get install -y openjdk-8-jdk RUN apt-get install -y vim RUN apt install -y net-tools RUN apt install -y iputils-ping RUN apt-get install -y openssh-server RUN mkdir /var/run/sshd RUN echo 'root:root' |chpasswd RUN sed -ri 's/^PermitRootLogin\s+.*/PermitRootLogin yes/' /etc/ssh/sshd_config RUN sed -ri 's/UsePAM yes/#UsePAM yes/g' /etc/ssh/sshd_config RUN sed -ri 's/# StrictHostKeyChecking ask/StrictHostKeyChecking no/' /etc/ssh/ssh_config RUN mkdir /root/.ssh RUN ssh-keygen -t rsa -P "" -f /root/.ssh/id_rsa RUN cat /root/.ssh/id_rsa.pub >> /root/.ssh/authorized_keys ENV JAVA_HOME /usr/lib/jvm/java-8-openjdk-amd64 ENV JRE_HOME /usr/lib/jvm/java-8-openjdk-amd64/jre ENV PATH /opt/hadoop-2.7.3/bin:/opt/hadoop-2.7.3/sbin:/usr/lib/jvm/java-8-openjdk-amd64/bin:$PATH ENV CLASSPATH ./:/usr/lib/jvm/java-8-openjdk-amd64/lib:/usr/lib/jvm/java-8-openjdk-amd64/jre/lib ADD hadoop-2.7.3.tar.gz /opt/ EXPOSE 22 CMD ["/usr/sbin/sshd", "-D"]