Spark 2.4 集群部署（on Yarn模式）

Posted 2021-04-27 小峰伊凡

基本信息

---

系统版本：3.10.0-1062.9.1.el7.x86_64

JDK 版本：1.8.0_202

---

Hadoop：https://mirrors.aliyun.com/apache/hadoop/common/hadoop-2.7.7/hadoop-2.7.7.tar.gzScala：https://github.com/scala/scala/releases/tag/v2.12.9 页面下下载tar.gz的包Spark：https://mirrors.aliyun.com/apache/spark/spark-2.4.7/spark-2.4.7-bin-hadoop2.7.tgz

Step 1：修改主机名称

hostnamectl set-hostname node01hostnamectl set-hostname node02hostnamectl set-hostname node03

Step 2：修改 hosts 文件

vim /etc/hosts

192.168.24.2 node01192.168.24.4 node02192.168.24.6 node03

Step 3：关闭防火墙，并禁止启动

systemctl stop firewalldsystemctl disable firewalld

Step 4：关闭 SELINUX

vim /etc/selinux/config将 SELINUX 配置项改为 SELINUX=disabled

Step 5：SSH 免密登录设置（所有节点）

ssh-keygen -t rsacat ~/.ssh/id_rsa.pub >> ~/.ssh/authorized_keyschmod 700 ~/.sshchmod 600 ~/.ssh/authorized_keys

# 将 node01 的公钥 copy 到其他节点，仅在 node01 上执行即可；

ssh-copy-id node02ssh-copy-id node03

Step 6：创建目录

mkdir -pv /home/hadoopmkdir -pv /home/hadoop/work/tmp/dfs/namemkdir -pv /home/hadoop/work/tmp/dfs/datamkdir -pv /home/spark

Step 7：下载 hadoop tar 包并解压

cd /home/hadoopwget https://mirrors.aliyun.com/apache/hadoop/common/hadoop-2.7.7/hadoop-2.7.7.tar.gztar zxvf hadoop-2.7.7.tar.gz

Step 8：进入目录 hadoop-2.7.7/etc/hadoop，依次编辑 hadoop-env.sh、mapred-env.sh、yarn-env.sh 这三个配置文件，确保内容中 JAVA_HOME 配置为正确路径，如下：

export JAVA_HOME=/usr/java/jdk1.8.0_202-amd64

Step 9：编辑 core-site.xml 文件，配置 configuration 节点，如下：

<configuration> <property> <name>fs.defaultFS</name> <value>hdfs://node01:8020</value> </property> <property> <name>hadoop.tmp.dir</name> <value>/home/hadoop/work/tmp</value> </property> <property> <name>dfs.namenode.name.dir</name> <value>file://${hadoop.tmp.dir}/dfs/name</value> </property> <property> <name>dfs.datanode.data.dir</name> <value>file://${hadoop.tmp.dir}/dfs/data</value> </property></configuration>

Step 10：编辑 hdfs-site.xml 文件，配置 configuration节点，将  node02 配置成 Sendary Namenode，如下：

<configuration> <property> <name>dfs.namenode.secondary.http-address</name> <value>node02:50090</value> </property></configuration>

Step 11：编辑 slaves 文件，如下：

node01node02node03

Step 12：编辑 yarn-site.xml 文件，配置 configuration 节点，如下：

<configuration><!-- Site specific YARN configuration properties --> <property> <name>yarn.nodemanager.aux-services</name> <value>mapreduce_shuffle</value> </property> <property> <name>yarn.resourcemanager.hostname</name> <value>node01</value> </property> <property> <name>yarn.log-aggregation-enable</name> <value>true</value> </property> <property> <name>yarn.log-aggregation.retain-seconds</name> <value>106800</value> </property> <property> <name>yarn.nodemanager.pmem-check-enabled</name> <value>false</value> </property> <property> <name>yarn.nodemanager.vmem-check-enabled</name> <value>false</value> </property></configuration>

Step 13：备份mapred-site.xml.template文件，将mapred-site.xml.template 重命名为 mapred-site.xml

cp mapred-site.xml.template mapred-site.xml.template.bakmv mapred-site.xml.template mapred-site.xml

Step 14：编辑 mapred-site.xml 文件，配置 configuration 节点，如下：

<configuration> <property> <name>mapreduce.framework.name</name> <value>yarn</value> </property> <property> <name>mapreduce.jobhistory.address</name> <value>node01:10020</value> </property> <property> <name>mapreduce.jobhistory.webapp.address</name> <value>node01:19888</value> </property></configuration>

Step 15：同步 hadoop-2.7.7 目前到其他节点

scp -r hadoop-2.7.7 node02:/home/hadoopscp -r hadoop-2.7.7 node03:/home/hadoop

Step 16：格式化 HDFS

cd /home/hadoop/hadoop-2.7.7bin/hdfs namenode -format

Step 17：启动 Hadoop

# 启动hdfs
sh /home/hadoop/hadoop-2.7.7/sbin/start-dfs.sh#启动yarn
sh /home/hadoop/hadoop-2.7.7/sbin/start-yarn.sh#启动日志服务
/home/hadoop/hadoop-2.7.7/sbin/mr-jobhistory-daemon.sh start historyserver

Step 18：下载 scala、spark 包并解压

cd /home/sparkhttps://github.com/scala/scala/releases/tag/v2.12.9 页面下载scala并将包上传至服务器 /home/spark目录；wget https://mirrors.aliyun.com/apache/spark/spark-2.4.7/spark-2.4.7-bin-hadoop2.7.tgz；
tar -zxvf scala-2.12.9.tar.gztar -zxvf spark-2.4.7-bin-hadoop2.7.tgz

Step 19：编辑 spark 相关配置文件

cd spark-2.4.7-bin-hadoop2.7/confcp spark-env.sh.template spark-env.sh.template.bakmv spark-env.sh.template spark-env.sh

vim spark-env.sh 追加：

export SPARK_MASTER_IP=node01export SPARK_MASTER_PORT=7077export SPARK_EXECUTOR_INSTANCES=1export SPARK_WORKER_INSTANCES=1export SPARK_WORKER_CORES=1export SPARK_WORKER_MEMORY=256Mexport SPARK_MASTER_WEBUI_PORT=8080export SPARK_CONF_DIR=/home/spark/spark-2.4.7-bin-hadoop2.7/confexport JAVA_HOME=/usr/java/jdk1.8.0_202-amd64export JRE_HOME=${JAVA_HOME}/jreexport HADOOP_CONF_DIR=/home/hadoop/hadoop-2.7.7/etc/hadoop

mv slaves.template slaves

vim slaves 

node01node02node03

Step 20：将 scala-2.12.9、spark-2.4.7-bin-hadoop2.7 目录同步到其他节点；

scp -r scala-2.12.9 node02:/home/sparkscp -r scala-2.12.9 node03:/home/sparkscp -r spark-2.4.7-bin-hadoop2.7 node02:/home/sparkscp -r spark-2.4.7-bin-hadoop2.7 node03:/home/spark

Step 21：启动 Spark

sh /home/spark/spark-2.4.7-bin-hadoop2.7/sbin/start-all.sh

Step 22：配置环境变量，并使之立即生效

#JAVAexport JAVA_HOME=/usr/java/jdk1.8.0_202-amd64export JRE_HOME=${JAVA_HOME}/jreexport CLASSPATH=.:${JAVA_HOME}/lib:${JRE_HOME}/libexport PATH=${JAVA_HOME}/bin:$PATH#SCALAexport SCALA_HOME=/home/spark/scala-2.12.9export PATH=${SCALA_HOME}/bin:$PATH#SPARKexport SPARK_HOME=/home/spark/spark-2.4.7-bin-hadoop2.7export PATH=${SPARK_HOME}/bin:$PATH#HADOOPexport HADOOP_HOME=/home/hadoop/hadoop-2.7.7export PATH=$HADOOP_HOME/bin:$HADOOP_HOME/sbin:$PATHexport HADOOP_CONF_DIR=/home/hadoop/hadoop-2.7.7/etc/hadoop

source /etc/profile

Step 23：简单配置一键启动脚本和关闭脚本方便启动与关闭

start-spark.sh #!/bin/bash
sh /home/hadoop/hadoop-2.7.7/sbin/start-dfs.sh \&& sh /home/hadoop/hadoop-2.7.7/sbin/start-yarn.sh \&& /home/hadoop/hadoop-2.7.7/sbin/mr-jobhistory-daemon.sh start historyserver \&& sh /home/spark/spark-2.4.7-bin-hadoop2.7/sbin/start-all.sh

stop-spark.sh #!/bin/bash
sh /home/spark/spark-2.4.7-bin-hadoop2.7/sbin/stop-all.sh \&& /home/hadoop/hadoop-2.7.7/sbin/mr-jobhistory-daemon.sh stop historyserver \&& sh /home/hadoop/hadoop-2.7.7/sbin/stop-yarn.sh \&& sh /home/hadoop/hadoop-2.7.7/sbin/stop-dfs.sh

Step 24：访问URL：http://192.168.24.2:8080