JavaWeb调用Hadoop进行MapReduce示例

Hadoop环境搭建见前面的帖子。

Hadoop下进行MapReduce见前面的帖子。

我们已经知道Hadoop能够通过Hadoop jar ***.jar input output的形式通过命令行来调用，那么如何将其封装成一个服务，让Java/Web来调用它？使得用户可以用方便的方式上传文件到Hadoop并进行处理，获得结果。首先，***.jar是一个Hadoop任务类的封装，我们可以在没有jar的情况下运行该类的main方法，将必要的参数传递给它。input 和output则将用户上传的文件使用Hadoop的JavaAPI put到Hadoop的文件系统中。然后再通过Hadoop的JavaAPI 从文件系统中取得结果文件。

搭建JavaWeb工程。本文使用Spring、SpringMVC、MyBatis框架， 当然，这不是重点，就算没有使用任何框架也能实现。

项目框架如下：

项目中使用到的jar包如下：

在Spring的配置文件中，加入

	<bean id="multipartResolver" class="org.springframework.web.multipart.commons.CommonsMultipartResolver">
		  <property name="defaultEncoding" value="utf-8" />
	      <property name="maxUploadSize" value="10485760000" />
	      <property name="maxInMemorySize" value="40960" />
	</bean>

使得项目支持文件上传。

新建一个login.jsp 点击登录后进入user/login

user/login中处理登录，登录成功后，【在Hadoop文件系统中创建用户文件夹】，然后跳转到console.jsp

package com.chenjie.controller;



import java.io.IOException;



import javax.annotation.Resource;

import javax.servlet.http.HttpServletRequest;

import javax.servlet.http.HttpServletResponse;



import org.apache.hadoop.conf.Configuration;

import org.apache.hadoop.fs.FileSystem;

import org.apache.hadoop.fs.Path;

import org.springframework.stereotype.Controller;

import org.springframework.web.bind.annotation.RequestMapping;



import com.chenjie.pojo.JsonResult;

import com.chenjie.pojo.User;

import com.chenjie.service.UserService;

import com.chenjie.util.AppConfig;

import com.google.gson.Gson;



/**

 * 用户请求控制器

 * 

 * @author Chen

 * 

 */

@Controller

// 声明当前类为控制器

@RequestMapping("/user")

// 声明当前类的路径

public class UserController {

	@Resource(name = "userService")

	private UserService userService;// 由Spring容器注入一个UserService实例



	/**

	 * 登录

	 * 

	 * @param user

	 *            用户

	 * @param request

	 * @param response

	 * @throws IOException

	 */

	@RequestMapping("/login")

	// 声明当前方法的路径

	public String login(User user, HttpServletRequest request,

			HttpServletResponse response) throws IOException {

		response.setContentType("application/json");// 设置响应内容格式为json

		User result = userService.login(user);// 调用UserService的登录方法

		request.getSession().setAttribute("user", result);

		if (result != null) {

			createHadoopFSFolder(result);

			return "console";

		}

		return "login";

	}



	


	public void createHadoopFSFolder(User user) throws IOException {

		Configuration conf = new Configuration();

		conf.addResource(new Path("/opt/hadoop-1.2.1/conf/core-site.xml"));

		conf.addResource(new Path("/opt/hadoop-1.2.1/conf/hdfs-site.xml"));



		FileSystem fileSystem = FileSystem.get(conf);

		System.out.println(fileSystem.getUri());



		Path file = new Path("/user/" + user.getU_username());

		if (fileSystem.exists(file)) {

			System.out.println("haddop hdfs user foler  exists.");

			fileSystem.delete(file, true);

			System.out.println("haddop hdfs user foler  delete success.");

		}

		fileSystem.mkdirs(file);

		System.out.println("haddop hdfs user foler  creat success.");

	}



}

console.jsp中进行文件上传和任务提交、

文件上传和任务提交:

package com.chenjie.controller;

import java.io.File;
import java.io.IOException;
import java.net.InetSocketAddress;
import java.net.URI;
import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;

import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FSDataInputStream;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.mapred.JobClient;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapred.JobID;
import org.apache.hadoop.mapred.JobStatus;
import org.apache.hadoop.mapred.RunningJob;
import org.springframework.stereotype.Controller;
import org.springframework.web.bind.annotation.RequestMapping;
import org.springframework.web.multipart.MultipartFile;
import org.springframework.web.multipart.MultipartHttpServletRequest;
import org.springframework.web.multipart.commons.CommonsMultipartResolver;

import com.chenjie.pojo.User;
import com.chenjie.util.Utils;

@Controller
// 声明当前类为控制器
@RequestMapping("/hadoop")
// 声明当前类的路径
public class HadoopController {

	@RequestMapping("/upload")
	// 声明当前方法的路径
	//文件上传
	public String upload(HttpServletRequest request,
			HttpServletResponse response) throws IOException {
		List<String> fileList = (List<String>) request.getSession()
				.getAttribute("fileList");//得到用户已上传文件列表
		if (fileList == null)
			fileList = new ArrayList<String>();//如果文件列表为空，则新建
		User user = (User) request.getSession().getAttribute("user");
		if (user == null)
			return "login";//如果用户未登录，则跳转登录页面
		CommonsMultipartResolver multipartResolver = new CommonsMultipartResolver(
				request.getSession().getServletContext());//得到在Spring配置文件中注入的文件上传组件
		if (multipartResolver.isMultipart(request)) {//如果请求是文件请求
			MultipartHttpServletRequest multiRequest = (MultipartHttpServletRequest) request;

			Iterator<String> iter = multiRequest.getFileNames();//得到文件名迭代器
			while (iter.hasNext()) {
				MultipartFile file = multiRequest.getFile((String) iter.next());
				if (file != null) {
					String fileName =  file.getOriginalFilename();
					File folder = new File("/home/chenjie/CJHadoopOnline/"
							+ user.getU_username());
					if (!folder.exists()) {
						folder.mkdir();//如果文件不目录存在，则在服务器本地创建
					}
					String path = "/home/chenjie/CJHadoopOnline/"
							+ user.getU_username() + "/" + fileName;

					File localFile = new File(path);

					file.transferTo(localFile);//将上传文件拷贝到服务器本地目录
					// fileList.add(path);
				}
				handleUploadFiles(user, fileList);//处理上传文件
			}

		}
		request.getSession().setAttribute("fileList", fileList);//将上传文件列表保存在Session中
		return "console";//返回console.jsp继续上传文件
	}

	@RequestMapping("/wordcount")
	//调用Hadoop进行mapreduce
	public void wordcount(HttpServletRequest request,
			HttpServletResponse response) {
		System.out.println("进入controller wordcount ");
		User user = (User) request.getSession().getAttribute("user");
		System.out.println(user);
		// if(user == null)
		// return "login";
		WordCount c = new WordCount();//新建单词统计任务
		String username = user.getU_username();
		String input = "hdfs://chenjie-virtual-machine:9000/user/" + username
				+ "/wordcountinput";//指定Hadoop文件系统的输入文件夹
		String output = "hdfs://chenjie-virtual-machine:9000/user/" + username
				+ "/wordcountoutput";//指定Hadoop文件系统的输出文件夹
		String reslt = output + "/part-r-00000";//默认输出文件
		try {
			Thread.sleep(3*1000);
			c.main(new String[] { input, output });//调用单词统计任务
			Configuration conf = new Configuration();//新建Hadoop配置
			conf.addResource(new Path("/opt/hadoop-1.2.1/conf/core-site.xml"));//添加Hadoop配置，找到Hadoop部署信息
			conf.addResource(new Path("/opt/hadoop-1.2.1/conf/hdfs-site.xml"));//Hadoop配置，找到文件系统

			FileSystem fileSystem = FileSystem.get(conf);//得打文件系统
			Path file = new Path(reslt);//找到输出结果文件
			FSDataInputStream inStream = fileSystem.open(file);//打开
			URI uri = file.toUri();//得到输出文件路径
			System.out.println(uri);
			String data = null;
			while ((data = inStream.readLine()) != null) {
				//System.out.println(data);
				response.getOutputStream().println(data);//讲结果文件写回用户网页
			}
//			InputStream in = fileSystem.open(file);
//			OutputStream out = new FileOutputStream("result.txt");
//			IOUtils.copyBytes(in, out, 4096, true);
			inStream.close();
		} catch (Exception e) {
			System.err.println(e.getMessage());
		}
	}

	@RequestMapping("/MapReduceStates")
	//得到MapReduce的状态
	public void mapreduce(HttpServletRequest request,
			HttpServletResponse response) {
		float[] progress=new float[2];
		try {
			Configuration conf1=new Configuration();
			conf1.set("mapred.job.tracker", Utils.JOBTRACKER);
			
			JobStatus jobStatus  = Utils.getJobStatus(conf1);
//			while(!jobStatus.isJobComplete()){
//				progress = Utils.getMapReduceProgess(jobStatus);
//				response.getOutputStream().println("map:" + progress[0]  + "reduce:" + progress[1]);
//				Thread.sleep(1000);
//			}
			JobConf jc = new JobConf(conf1);
			
			JobClient jobClient = new JobClient(jc);
			JobStatus[] jobsStatus = jobClient.getAllJobs();  
			//这样就得到了一个JobStatus数组，随便取出一个元素取名叫jobStatus  
			jobStatus = jobsStatus[0];  
			JobID jobID = jobStatus.getJobID(); //通过JobStatus获取JobID  
			RunningJob runningJob = jobClient.getJob(jobID);  //通过JobID得到RunningJob对象  
			runningJob.getJobState();//可以获取作业状态，状态有五种，为JobStatus.Failed 、JobStatus.KILLED、JobStatus.PREP、JobStatus.RUNNING、JobStatus.SUCCEEDED  
			jobStatus.getUsername();//可以获取运行作业的用户名。  
			runningJob.getJobName();//可以获取作业名。  
			jobStatus.getStartTime();//可以获取作业的开始时间，为UTC毫秒数。  
			float map = runningJob.mapProgress();//可以获取Map阶段完成的比例，0~1，  
			System.out.println("map=" + map);
			float reduce = runningJob.reduceProgress();//可以获取Reduce阶段完成的比例。
			System.out.println("reduce="+reduce);
			runningJob.getFailureInfo();//可以获取失败信息。  
			runningJob.getCounters();//可以获取作业相关的计数器，计数器的内容和作业监控页面上看到的计数器的值一样。 
			
			
		} catch (IOException e) {
			progress[0] = 0;
			progress[1] = 0;
		}
	
		request.getSession().setAttribute("map", progress[0]);
		request.getSession().setAttribute("reduce", progress[1]);
	}
	
	//处理文件上传
	public void handleUploadFiles(User user, List<String> fileList) {
		File folder = new File("/home/chenjie/CJHadoopOnline/"
				+ user.getU_username());
		if (!folder.exists())
			return;
		if (folder.isDirectory()) {
			File[] files = folder.listFiles();
			for (File file : files) {
				System.out.println(file.getName());
				try {
					putFileToHadoopFSFolder(user, file, fileList);//将单个文件上传到Hadoop文件系统
				} catch (IOException e) {
					System.err.println(e.getMessage());
				}
			}
		}
	}

	//将单个文件上传到Hadoop文件系统
	private void putFileToHadoopFSFolder(User user, File file,
			List<String> fileList) throws IOException {
		Configuration conf = new Configuration();
		conf.addResource(new Path("/opt/hadoop-1.2.1/conf/core-site.xml"));
		conf.addResource(new Path("/opt/hadoop-1.2.1/conf/hdfs-site.xml"));

		FileSystem fileSystem = FileSystem.get(conf);
		System.out.println(fileSystem.getUri());

		Path localFile = new Path(file.getAbsolutePath());
		Path foler = new Path("/user/" + user.getU_username()
				+ "/wordcountinput");
		if (!fileSystem.exists(foler)) {
			fileSystem.mkdirs(foler);
		}
		
		Path hadoopFile = new Path("/user/" + user.getU_username()
				+ "/wordcountinput/" + file.getName());
//		if (fileSystem.exists(hadoopFile)) {
//			System.out.println("File exists.");
//		} else {
//			fileSystem.mkdirs(hadoopFile);
//		}
		fileSystem.copyFromLocalFile(true, true, localFile, hadoopFile);
		fileList.add(hadoopFile.toUri().toString());

	}

}

启动Hadoop：

运行结果：

可以在任意平台下，登录该项目地址，上传文件，得到结果。

运行成功。

源代码：

https://github.com/tudoupaisimalingshu/CJHadoopOnline