1、环境准备：
下载spark-2.4.4-bin-hadoop2.7.tgz
上传安装包到Linux中
解压安装包
tar -zxf spark-2.4.4-bin-hadoop2.7.tgz -C /hadoop/app
2、配置环境
2.1修改hadoop中yarn-site.xml
<property>
<name>yarn.nodemanager.pmem-check-enabled</name>
<value>false</value>
</property>
<property>
<name>yarn.nodemanager.vmem-check-enabled</name>
<value>false</value>
</property>
将文件同步到其他节点
scp yarn-site,xml node2:/hadoop/app/hadoop-2.7.2/etc/hadoop/
scp yarn-site,xml node3:/hadoop/app/hadoop-2.7.2/etc/hadoop/
2.2修改spark配置文件
cd /hadoop/app/spark-2.4.4-bin-hadoop2.7/conf
vim slaves
vim spark-env.sh
同步配置文件到其他节点
scp slaves node2:/hadoop/app/spark-2.4.4-bin-hadoop2.7/conf
scp slaves node3:/hadoop/app/spark-2.4.4-bin-hadoop2.7/conf
scp spark-env.sh node2:/hadoop/app/spark-2.4.4-bin-hadoop2.7/conf
scp spark-env.sh node3:/hadoop/app/spark-2.4.4-bin-hadoop2.7/conf
3.环境验证
3.1启动yarn
spark-submit --class org.apache.spark.examples.SparkPi --master yarn --deploy-mode client /hadoop/app/spark-2.4.4-bin-hadoop2.7examples/jars/spark-examples_2.11-2.4.4.jar 100