一、解压压缩包

[hadoop@hadoop102 software]$ tar -zxvf spark-3.0.0-bin-hadoop3.2.tgz -C /opt/module

二、修改配置文件conf/spark-env.sh

cd /opt/module/spark-3.0.0-bin-hadoop3.2/conf
[hadoop@hadoop102 conf]$ cp spark-env.sh.template spark-env.sh
[hadoop@hadoop102 conf]$ vim spark-env.sh

内容：

export JAVA_HOME=/opt/module/jdk1.8.0_291
YARN_CONF_DIR=/opt/module/hadoop-3.1.3/etc/hadoop

三、测试提交Spark任务

[hadoop@hadoop102 ~]$ cd /opt/module/spark-3.0.0-bin-hadoop3.2/
[hadoop@hadoop102 spark-3.0.0-bin-hadoop3.2]$ bin/spark-submit \
> --class org.apache.spark.examples.SparkPi \
> --master yarn \
> --deploy-mode cluster \
> ./examples/jars/spark-examples_2.12-3.0.0.jar \
> 10

到YARN WEB页面查看任务提交情况

四、Spark on Hive配置

4.1 创建hive-site.xml（spark/conf目录）

[hadoop@hadoop102 conf]$ cd /opt/module/spark-3.0.0-bin-hadoop3.2/conf/
[hadoop@hadoop102 conf]$ vim hive-site.xml

内容：

<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
<configuration><!--告知Spark创建表存到哪里--><property><name>hive.metastore.warehouse.dir</name><value>/warehouse</value></property><!-- 不使用spark内置hive存储元数据 --><property><name>hive.metastore.local</name><value>false</value></property><!--告知Spark Hive的MetaStore在哪--><property><name>hive.metastore.uris</name><value>thrift://hadoop102:9083</value></property></configuration>

4.2 查看hive的hive-site.xml配置与3.1配置的是否一致

cd /opt/module/apache-hive-3.1.2-bin/conf
vim hive-site.xml

4.3 测试SparkSQL

4.3.1 启动SparkSQL客户端

[hadoop@hadoop102 spark-3.0.0-bin-hadoop3.2]$ bin/spark-sqlshow databases;select count(1)
from dw_ods.ods_activity_info_full 
where dt='2023-12-07';

4.3.2 启动Hive客户端

[hadoop@hadoop102 apache-hive-3.1.2-bin]$ bin/hiveshow databases;select count(1)
from dw_ods.ods_activity_info_full 
where dt='2023-12-07';

五、集群化（TODO）

优势在哪里？？