一、安装Scala
解压Scala压缩包到soft目录，并改名为scala211
cd /opt/soft/install/
ls
tar -zxvf scala-2.11.12.tgz -C /opt/soft
cd /opt/soft
ls
mv scala-2.11.12 scala211
然后，修改一下profile文件
vi /etc/profile
添加spark安装目录
#scala
export SCALA_HOME=/opt/soft/scala211
export PATH=$PATH:$SCALA_HOME/bin
然后，source一下
source /etc/profile
测试一下
Scala
二、安装spark
cd /opt/soft/install/
ls
tar -zxvf spark-2.4.5-bin-hadoop2.6.tgz -C /opt/soft
cd /opt/soft
ls
mv spark-2.4.5-bin-hadoop2.6 spark245
ls
cd /opt/soft/spark245/
ls
cd /opt/soft/spark245/
pwd
vi /etc/profile
添加spark安装目录
#spark
export SPARK_HOME=/opt/soft/spark245
export PATH=$PATH:$SPARK_HOME/bin
cd ./conf/
ls
cp spark-env.sh.template spark-env.sh
ls
vi ./spark-env.sh
export JAVA_HOME=/opt/jdk1.8.0_221
export SCALA_HOME=/opt/soft/scala211
export SPARK_HOME=/opt/soft/spark245
export HADOOP_INSTALL=/opt/soft/install/hadoop260
export HADOOP_CONF_DIR=$HADOOP_INSTALL/etc/hadoop
export SPARK_MASTER_IP=gree128
export SPARK_DRIVER_MEMORY=2G
export SPARK_EXECUTOR_MEMORY=2G
export SPARK_LOCAL_DIRS=/opt/soft/spark245
cp slaves.template slaves
vi ./slaves
localhost
然后，source一下
source /etc/profile
然后，先启动Hadoop，再以yarn方式启动spark
spark-shell
1.本机 spark-shell --master local[*]
2.Standalone（小集群） spark-shell --master spark://MASTERHOST:7077
3.YARN spark-shell --master yarn-client
