开源大数据集群部署（二十一）Spark on yarn 部署

作者：櫰木

1 spark on yarn安装（每个节点）

cd /root/bigdata/
tar -xzvf spark-3.3.1-bin-hadoop3.tgz -C /opt/
ln -s /opt/spark-3.3.1-bin-hadoop3 /opt/spark
chown -R spark:spark /opt/spark-3.3.1-bin-hadoop3

2 配置环境变量及修改配置

cat /etc/profile.d/bigdata.sh
export SPARK_HOME=/opt/spark
export SPARK_CONF_DIR=/opt/spark/conf

引用变量

source /etc/profile

yarn的capacity-scheduler.xml文件修改配置保证资源调度按照CPU + 内存模式：(每个yarn 节点)

<property>
    <name>yarn.scheduler.capacity.resource-calculator</name>
    <!-- <value>org.apache.hadoop.yarn.util.resource.DefaultResourceCalculator</value> -->
    <value>org.apache.hadoop.yarn.util.resource.DominantResourceCalculator</value>
</property>

在yarn-site.xml开启日志功能：

<property>
    <description>Whether to enable log aggregation</description>
    <name>yarn.log-aggregation-enable</name>
    <value>true</value>
</property>
<property>
    <name>yarn.log.server.url</name>
    <value>http://master:19888/jobhistory/logs</value>
</property>

修改mapred-site.xml: （每个yarn节点）

 <property>
        <name>mapreduce.jobhistory.address</name>
        <value>hd1.dtstack.com:10020</value>
    </property>
    <property>
        <name>mapreduce.jobhistory.webapp.address</name>
        <value>hd1.dtstack.com:19888</value>
</property>

cd /opt/spark/conf

Spark 配置文件（每个spark节点）

cat spark-defaults.conf
spark.eventLog.dir=hdfs:///user/spark/applicationHistory
spark.eventLog.enabled=true
spark.yarn.historyServer.address=http://hd1.dtstack.com:18018
 
spark.history.kerberos.enabled=true
 
spark.history.kerberos.principal=hdfs/hd1.dtstack.com@DTSTACK.COM
spark.history.kerberos.keytab=/etc/security/keytab/hdfs.keytab

Spark 环境配置文件（每个spark节点）

cat spark-env.sh
export HADOOP_CONF_DIR=${HADOOP_HOME}/etc/hadoop
export SPARK_HISTORY_OPTS="-Dspark.history.ui.port=18018 -Dspark.history.fs.logDirectory=hdfs:///user/spark/applicationHistory"
export YARN_CONF_DIR=${HADOOP_HOME}/etc/hadoop

由于需要读取日志文件，所以使用hdfs的keytab

创建对应hdfs目录，并修改权限

hdfs dfs -mkdir -p /user/spark/applicationHistory
hdfs dfs -chown -R spark /user/spark/

提交测试任务

cd /opt/spark
./bin/spark-submit --master yarn --deploy-mode client --class org.apache.spark.examples.SparkPi examples/jars/spark-examples_2.12-3.3.1.jar