java sparksubmit_提交任务到Spark

1.场景

在搭建好Hadoop+Spark环境后，现准备在此环境上提交简单的任务到Spark进行计算并输出结果。搭建过程：http://www.cnblogs.com/zengxiaoliang/p/6478859.html

本人比较熟悉Java语言，现以Java的WordCount为例讲解这整个过程，要实现计算出给定文本中每个单词出现的次数。

2.环境测试

在讲解例子之前，我想先测试一下之前搭建好的环境。

2.1测试Hadoop环境

首先创建一个文件wordcount.txt 内容如下：

Hello hadoop

hello spark

hello bigdata

yellow banana

red apple

然后执行如下命令：

hadoop fs -mkdir -p /Hadoop/Input(在HDFS创建目录)

hadoop fs -put wordcount.txt /Hadoop/Input(将wordcount.txt文件上传到HDFS)

hadoop fs -ls /Hadoop/Input (查看上传的文件)

hadoop fs -text /Hadoop/Input/wordcount.txt (查看文件内容)

2.2Spark环境测试

我使用spark-shell，做一个简单的WordCount的测试。我就用上面Hadoop测试上传到HDFS的文件wordcount.txt。

首先启动spark-shell命令：

spark-shell

然后直接输入scala语句：

val file=sc.textFile("hdfs://Master:9000/Hadoop/Input/wordcount.txt")

val rdd = file.flatMap(line => line.split(" ")).map(word => (word,1)).reduceByKey(_+_)

rdd.collect()

rdd.foreach(println)

退出使用如下命令：

:quit

这样环境测试就结束了。

3.Java实现单词计数

packagecom.example.spark;importjava.util.Arrays;importjava.util.Iterator;importjava.util.List;importjava.util.regex.Pattern;importorg.apache.spark.SparkConf;importorg.apache.spark.api.java.JavaPairRDD;importorg.apache.spark.api.java.JavaRDD;importorg.apache.spark.api.java.JavaSparkContext;importorg.apache.spark.api.java.function.FlatMapFunction;importorg.apache.spark.api.java.function.Function2;importorg.apache.spark.api.java.function.PairFunction;importscala.Tuple2;public final classWordCount {private static final Pattern SPACE = Pattern.compile(" ");public static void main(String[] args) throwsException {

SparkConf conf= new SparkConf().setAppName("kevin's first spark app");

JavaSparkContext sc= newJavaSparkContext(conf);

JavaRDD lines = sc.textFile(args[0]).cache();

JavaRDD words = lines.flatMap(new FlatMapFunction() {private static final long serialVersionUID = 1L;

@Overridepublic Iteratorcall(String s) {returnArrays.asList(SPACE.split(s)).iterator();

}

});

JavaPairRDD ones = words.mapToPair(new PairFunction() {private static final long serialVersionUID = 1L;

@Overridepublic Tuple2call(String s) {return new Tuple2(s, 1);

}

});

JavaPairRDD counts = ones.reduceByKey(new Function2() {private static final long serialVersionUID = 1L;

@OverridepublicInteger call(Integer i1, Integer i2) {return i1 +i2;

}

});

List> output =counts.collect();for (Tuple2, ?>tuple : output) {

System.out.println(tuple._1()+ ": " +tuple._2());

}

sc.close();

}

}

4.任务提交实现

将上面Java实现的单词计数打成jar包spark-example-0.0.1-SNAPSHOT.jar，并且将jar包上传到Master节点，我是将jar包上传到/opt目录下，本文将以两种方式提交任务到spark，第一种是以spark-submit命令的方式提交任务，第二种是以java web的方式提交任务。

4.1以spark-submit命令的方式提交任务

spark-submit --master spark://114.55.246.88:7077 --class com.example.spark.WordCount /opt/spark-example-0.0.1-SNAPSHOT.jar hdfs://Master:9000/Hadoop/Input/wordcount.txt

4.2以java web的方式提交任务

我是用spring boot搭建的java web框架，实现代码如下：

1)新建maven项目spark-submit

2)pom.xml文件内容，这里要注意spark的依赖jar包要与scala的版本相对应，如spark-core_2.11，这后面2.11就是你安装的scala的版本。

4.0.0

org.springframework.boot

spring-boot-starter-parent

1.4.1.RELEASE

spark-submit

spark-submit

com.example.spark.SparkSubmitApplication

UTF-8

1.8

3.4

2.1.0

org.apache.commons

commons-lang3

${commons.version}

org.apache.tomcat.embed

tomcat-embed-jasper

provided

org.springframework.boot

spring-boot-starter-data-jpa

org.springframework.boot

spring-boot-starter-data-redis

org.springframework.boot

spring-boot-starter-test

test

com.jayway.jsonpath

json-path

org.springframework.boot

spring-boot-starter-web

spring-boot-starter-tomcat

org.springframework.boot

org.springframework.boot

spring-boot-starter-jetty

org.eclipse.jetty.websocket

*

org.springframework.boot

spring-boot-starter-jetty

provided

javax.servlet

jstl

org.eclipse.jetty

apache-jsp

provided

org.springframework.boot

spring-boot-starter-data-solr

org.springframework.boot

spring-boot-starter-data-jpa

org.springframework.boot

spring-boot-starter-web

javax.servlet

jstl

org.apache.spark

spark-core_2.11

${org.apache.spark-version}

org.apache.spark

spark-sql_2.11

${org.apache.spark-version}

org.apache.spark

spark-hive_2.11

${org.apache.spark-version}

org.apache.spark

spark-streaming_2.11

${org.apache.spark-version}

org.apache.hadoop

hadoop-client

2.7.3

org.apache.spark

spark-streaming-kafka_2.11

1.6.3

org.apache.spark

spark-graphx_2.11

${org.apache.spark-version}

org.apache.maven.plugins

maven-assembly-plugin

3.0.0

com.fasterxml.jackson.core

jackson-core

2.6.5

com.fasterxml.jackson.core

jackson-databind

2.6.5

com.fasterxml.jackson.core

jackson-annotations

2.6.5

war

spring-snapshots

Spring Snapshots

https://repo.spring.io/snapshot

true

spring-milestones

Spring Milestones

https://repo.spring.io/milestone

false

maven2

http://repo1.maven.org/maven2/

spring-snapshots

Spring Snapshots

https://repo.spring.io/snapshot

true

spring-milestones

Spring Milestones

https://repo.spring.io/milestone

false

maven-war-plugin

src/main/webapp

org.mortbay.jetty

jetty-maven-plugin

spring.profiles.active

development

org.eclipse.jetty.server.Request.maxFormContentSize

600000

true

/

7080

3)SubmitJobToSpark.java

packagecom.example.spark;importorg.apache.spark.deploy.SparkSubmit;/***@authorkevin

**/

public classSubmitJobToSpark {public static voidsubmitJob() {

String[] args= new String[] { "--master", "spark://114.55.246.88:7077", "--name", "test java submit job to spark", "--class", "com.example.spark.WordCount", "/opt/spark-example-0.0.1-SNAPSHOT.jar", "hdfs://Master:9000/Hadoop/Input/wordcount.txt"};

SparkSubmit.main(args);

}

}

4)SparkController.java

packagecom.example.spark.web.controller;importjavax.servlet.http.HttpServletRequest;importjavax.servlet.http.HttpServletResponse;importorg.slf4j.Logger;importorg.slf4j.LoggerFactory;importorg.springframework.stereotype.Controller;importorg.springframework.web.bind.annotation.RequestMapping;importorg.springframework.web.bind.annotation.RequestMethod;importorg.springframework.web.bind.annotation.ResponseBody;importcom.example.spark.SubmitJobToSpark;

@Controller

@RequestMapping("spark")public classSparkController {private Logger logger = LoggerFactory.getLogger(SparkController.class);

@RequestMapping(value= "sparkSubmit", method ={ RequestMethod.GET, RequestMethod.POST })

@ResponseBodypublicString sparkSubmit(HttpServletRequest request, HttpServletResponse response) {

logger.info("start submit spark tast...");

SubmitJobToSpark.submitJob();return "hello";

}

}

5)将项目spark-submit打成war包部署到Master节点tomcat上，访问如下请求：

http://114.55.246.88:9090/spark-submit/spark/sparkSubmit

在tomcat的log中能看到计算的结果。