03. 搭建Spark集群(CentOS7+Spark2.1.1+Hadoop2.8.0)

Posted 2021-01-25 yjm0330

一、下载安装scala

1、官网下载

2、spar01和02都建立/opt/scala目录，解压tar -zxvf scala-2.12.8.tgz

3、配置环境变量

　　vi /etc/profile 增加一行

　　export    SCALA_HOME=/opt/scala/scala-2.12.8

　　同时把hadoop的环境变量增加进去，完整版是：

export JAVA_HOME=/opt/java/jdk1.8.0_191
export HADOOP_HOME=/opt/hadoop/hadoop-2.8.0
export HADOOP_CONF_DIR=${HADOOP_HOME}/etc/hadoop
export HADOOP_COMMON_LIB_NATIVE_DIR=${HADOOP_HOME}/lib/native
export HADOOP_OPTS="-Djava.library.path=${HADOOP_HOME}/lib"
export SCALA_HOME=/opt/scala/scala-2.12.8

export CLASSPATH=$:CLASSPATH:${JAVA_HOME}/lib/
export PATH=.:${JAVA_HOME}/bin:${HADOOP_HOME}/bin:${HADOOP_HOME}/sbin:${SPARK_HOME}/bin:${SCALA_HOME}/bin:$PATH

　　然后source /etc/profile

4、验证

　　scala -version

5、同步spark02配置文件

　　scp /etc/profile spark02:/etc

二、下载安装spark

1、下载，解压，同scala，建立/opt/spark目录

2、配置环境变量

export SPARK_HOME=/opt/spark/spark-2.4.0-bin-hadoop2.7

完整版更新：

export JAVA_HOME=/opt/java/jdk1.8.0_191
export HADOOP_HOME=/opt/hadoop/hadoop-2.8.0
export HADOOP_CONF_DIR=${HADOOP_HOME}/etc/hadoop
export HADOOP_COMMON_LIB_NATIVE_DIR=${HADOOP_HOME}/lib/native
export HADOOP_OPTS="-Djava.library.path=${HADOOP_HOME}/lib"
export SCALA_HOME=/opt/scala/scala-2.12.8
export SPARK_HOME=/opt/spark/spark-2.4.0-bin-hadoop2.7

export CLASSPATH=$:CLASSPATH:${JAVA_HOME}/lib/
export PATH=.:${JAVA_HOME}/bin:${HADOOP_HOME}/bin:${HADOOP_HOME}/sbin:${SPARK_HOME}/bin:${SCALA_HOME}/bin:$PATH

source /etc/profile

scp /etc/profile spark02:/etc

3、配置conf下文件

cp spark-env.sh.template spark-env.sh

cp slaves.template slaves

vi spark-env.sh

export SCALA_HOME=/opt/scala/scala-2.12.8
export JAVA_HOME=/opt/java/jdk1.8.0_191
export HADOOP_HOME=/opt/hadoop/hadoop-2.8.0
export HADOOP_CONF_DIR=$HADOOP_HOME/etc/hadoop
export SPARK_HOME=/opt/spark/spark-2.4.0-bin-hadoop2.7
export SPARK_MASTER_IP=spark01
export SPARK_EXECUTOR_MEMORY=2G

vi slaves

spark02

同步到spark02

scp /opt/spark/spark-2.4.0-bin-hadoop2.7/conf/spark-env.sh spark02:/opt/spark/spark-2.4.0-bin-hadoop2.7/conf/
scp /opt/spark/spark-2.4.0-bin-hadoop2.7/conf/slaves spark02:/opt/spark/spark-2.4.0-bin-hadoop2.7/conf/

三、测试spark

　　因为spark是依赖于hadoop提供的分布式文件系统的，所以在启动spark之前，先确保hadoop在正常运行。

　　在hadoop正常运行的情况下，在spark01（也就是hadoop的namenode，spark的marster节点）上执行命令：

　　cd   /opt/spark/spark-2.4.0-bin-hadoop2.7/sbin

　　执行启动脚本：./start-all.sh

　　在浏览器里访问Mster机器，我的Spark集群里Master机器是spark01，IP地址是192.168.2.245，访问8080端口，URL是：http://192.168.2.245:8080/

 

　　用local模式运行一个计算圆周率的Demo。按照下面的步骤来操作。

　　第一步，进入到Spark的根目录，也就是执行下面的脚本：

　　./bin/spark-submit  --class  org.apache.spark.examples.SparkPi  --master local   examples/jars/spark-examples_2.11-2.4.0.jar

　　yarn-client模式：

　　./bin/spark-submit  --class  org.apache.spark.examples.SparkPi  --master  yarn-client    examples/jars/spark-examples_2.11-2.4.0.jar