docker使用Dockerfile搭建spark集群

1、创建Dockerfile文件，内容如下

# 基础镜像，包括jdk
FROM openjdk:8u131-jre-alpine

#作者
LABEL maintainer "tony@163.com"

#用户
USER root

#编码

ENV LANG=C.UTF-8 
    TZ=Asia/Shanghai

#下载到时候安装spark需要的工具

RUN apk add --no-cache --update-cache bash curl tzdata wget tar 
    && cp /usr/share/zoneinfo/$TZ /etc/localtime 
    && echo $TZ > /etc/timezone

#设置工作目录

WORKDIR /usr/local

#拷贝 当下载过慢时先下载到本地在拷贝
# COPY spark-2.4.0-bin-hadoop2.7.tgz /usr/local 

#下载spark

RUN wget "http://archive.apache.org/dist/spark/spark-2.4.0/spark-2.4.0-bin-hadoop2.7.tgz" 
    && tar -vxf spark-* 
    && mv spark-2.4.0-bin-hadoop2.7 spark 
    && rm -rf spark-2.4.0-bin-hadoop2.7.tgz

#设定spark的home

ENV SPARK_HOME=/usr/local/spark 
    JAVA_HOME=/usr/lib/jvm/java-1.8-openjdk 
    PATH=${PATH}:${JAVA_HOME}/bin:${SPARK_HOME}/bin

#暴露端口，8080 6066 7077 4044 18080

EXPOSE 6066 8080 7077 4044 18080

#工作目录

WORKDIR $SPARK_HOME

CMD ["/bin/bash"]

2、在Dockerfile所在目录下构建镜像

docker build -t spark:v2.4.0 .

3、启动主节点

docker run -itd --name spark-master -p 6066:6066 -p 7077:7077 -p 8081:8080 -h spark-master spark:v2.4.0 ./bin/spark-class org.apache.spark.deploy.master.Master

4、启动从节点

docker run -itd --name spark-worker -P -h spark-worker --link spark-master spark:v2.4.0 ./bin/spark-class org.apache.spark.deploy.worker.Worker spark://spark-master:7077

5、浏览器访问

http://192.168.75.142:8081/

6、启动spark-shell客户端

docker exec -it 99306051a9d4 ./bin/spark-shell

其中 99306051a9d4 是 spark-master 的容器ID

7、spark-shell测试

scala> val file = sc.textFile("/usr/local/spark/bin/beeline")
file: org.apache.spark.rdd.RDD[String] = /usr/local/spark/bin/beeline MapPartitionsRDD[1] at textFile at <console>:24

scala> val words = file.flatMap(line => line.split(" ")).map(word => (word,1)).reduceByKey((a,b) => a+b)
words: org.apache.spark.rdd.RDD[(String, Int)] = ShuffledRDD[4] at reduceByKey at <console>:25

scala> words.collect
res0: Array[(String, Int)] = Array(($CLASS,1), (Software,1), (Unless,1), (this,3), (starting,1), (under,4), (The,1), (limitations,1),
("$0")"/find-spark-home,1), (Figure,1), (express,1), (contributor,1), (#,20), (WITHOUT,1), ("AS,1), (#!/usr/bin/env,1), (See,2), (License.,2), 
(for,4), (fi,1), (software,1), (IS",1), (obtain,1), (ANY,1), (SPARK_HOME,1), (out,1), (required,1), (2.0,1), (OR,1), (file,3), (the,9), (-o,1), 
(licenses,1), (not,1), (either,1), (if,2), (posix,2), (source,1), (Apache,2), (then,1), ("License");,1), (language,1), (License,3), (Enter,1),
(permissions,1), (WARRANTIES,1), (license,1), (by,1), (];,1), ("$(dirname,1), (an,1), ([,1), (agreed,1), (Version,1), (implied.,1), (KIND,,1), 
(is,2), ((the,1), (exec,1), ("${SPARK_HOME}",1), (agreements.,1), (on,1), (You,2), (one,1)...
scala>