为Hadoop的MapReduce程序编写makefile

最近需要把基于hadoop的MapReduce程序集成到一个大的用C/C++编写的框架中，需要在make的时候自动将MapReduce应用进行编译和打包。这里以简单的WordCount1为例说明具体的实现细节，注意：hadoop版本为2.4.0.

源代码包含两个文件，一个是WordCount1.java是具体的对单词计数实现的逻辑；第二个是CounterThread.java，其中简单的当前处理的行数做一个统计和打印。代码分别见附1. 编写makefile的关键是将hadoop提供的jar包的路径全部加载进来，看到网上很多资料都自己实现一个脚本把hadoop目录下所有的.jar文件放到一个路径中，然后进行编译，这种做法太麻烦了。当然也有些简单的办法，但是都是比较老的hadoop版本如0.20之类的。

其实，hadoop提供了一个命令hadoop classpath可以获得包含所有jar包的路径.所以只需要用 javac -classpath "`hadoop classpath`" *.java 便可，然后使用jar -cvf对class文件进行打包就可以了。具体的Makefile代码如下：

SRC_DIR = src/mypackage/*.java
CLASS_DIR = bin
TARGET_JAR = WordCount

all:$(TARGET_JAR)

$(TARGET_JAR): $(SRC_DIR)
mkdir -p $(CLASS_DIR)
#	javac -classpath `$(HADOOP) classpath` -d $(CLASS_DIR) $(SRC_DIR)
javac -classpath "`hadoop classpath`" src/mypackage/*.java -d $(CLASS_DIR) -Xlint
jar -cvf $(TARGET_JAR).jar -C $(CLASS_DIR) ./

clean:
rm -rf $(CLASS_DIR) *.jar

make一下：

lichao@ubuntu:WordCount1$ make
mkdir -p bin
javac -classpath "`hadoop classpath`" src/mypackage/*.java -d bin -Xlint
warning: [path] bad path element "/home/lichao/Software/hadoop/hadoop-src/hadoop-2.4.0-src/hadoop-dist/target/hadoop-2.4.0/share/hadoop/common/lib/jaxb-api.jar": no such file or directory
warning: [path] bad path element "/home/lichao/Software/hadoop/hadoop-src/hadoop-2.4.0-src/hadoop-dist/target/hadoop-2.4.0/share/hadoop/common/lib/activation.jar": no such file or directory
warning: [path] bad path element "/home/lichao/Software/hadoop/hadoop-src/hadoop-2.4.0-src/hadoop-dist/target/hadoop-2.4.0/share/hadoop/common/lib/jsr173_1.0_api.jar": no such file or directory
warning: [path] bad path element "/home/lichao/Software/hadoop/hadoop-src/hadoop-2.4.0-src/hadoop-dist/target/hadoop-2.4.0/share/hadoop/common/lib/jaxb1-impl.jar": no such file or directory
warning: [path] bad path element "/home/lichao/Software/hadoop/hadoop-src/hadoop-2.4.0-src/hadoop-dist/target/hadoop-2.4.0/share/hadoop/yarn/lib/jaxb-api.jar": no such file or directory
warning: [path] bad path element "/home/lichao/Software/hadoop/hadoop-src/hadoop-2.4.0-src/hadoop-dist/target/hadoop-2.4.0/share/hadoop/yarn/lib/activation.jar": no such file or directory
warning: [path] bad path element "/home/lichao/Software/hadoop/hadoop-src/hadoop-2.4.0-src/hadoop-dist/target/hadoop-2.4.0/share/hadoop/yarn/lib/jsr173_1.0_api.jar": no such file or directory
warning: [path] bad path element "/home/lichao/Software/hadoop/hadoop-src/hadoop-2.4.0-src/hadoop-dist/target/hadoop-2.4.0/share/hadoop/yarn/lib/jaxb1-impl.jar": no such file or directory
warning: [path] bad path element "/home/lichao/Software/hadoop/hadoop-src/hadoop-2.4.0-src/hadoop-dist/target/hadoop-2.4.0/contrib/capacity-scheduler/*.jar": no such file or directory
src/mypackage/WordCount1.java:61: warning: [deprecation] Job(Configuration,String) in Job has been deprecated
Job job = new Job(conf, "WordCount1");                  //建立新job
^
10 warnings
jar -cvf WordCount.jar -C bin ./
added manifest
adding: mypackage/(in = 0) (out= 0)(stored 0%)
adding: mypackage/WordCount1.class(in = 1970) (out= 1037)(deflated 47%)
adding: mypackage/CounterThread.class(in = 1760) (out= 914)(deflated 48%)
adding: mypackage/WordCount1$IntSumReducer.class(in = 1762) (out= 749)(deflated 57%)
adding: mypackage/WordCount1$TokenizerMapper.class(in = 1759) (out= 762)(deflated 56%)
adding: log4j.properties(in = 476) (out= 172)(deflated 63%)

虽然有warning，但是不影响结果。编译后，我们来简单的测试一下。

先生成测试数据：while true; do seq 1 100000 >> tmpfile; done; 差不多可以了就Ctrl+c

然后将数据放到hdfs上，hadoop fs -put tmpfile /data/

接着运行MapReduce程序：hadoop jar WordCount.jar mypackage/WordCount1 /data/tmpfile /output2

效果如下：

14/07/15 13:26:01 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
14/07/15 13:26:03 INFO client.RMProxy: Connecting to ResourceManager at localhost/127.0.0.1:8032
14/07/15 13:26:05 INFO input.FileInputFormat: Total input paths to process : 1
14/07/15 13:26:05 INFO mapreduce.JobSubmitter: number of splits:6
14/07/15 13:26:06 INFO mapreduce.JobSubmitter: Submitting tokens for job: job_1405397597558_0003
14/07/15 13:26:06 INFO impl.YarnClientImpl: Submitted application application_1405397597558_0003
14/07/15 13:26:06 INFO mapreduce.Job: The url to track the job: http://ubuntu:8088/proxy/application_1405397597558_0003/ 14/07/15 13:26:06 INFO mapreduce.Job: Running job: job_1405397597558_0003
14/07/15 13:26:20 INFO mapreduce.Job: Job job_1405397597558_0003 running in uber mode : false
14/07/15 13:26:20 INFO mapreduce.Job:  map 0% reduce 0%
14/07/15 13:26:34 WARN mapreduce.Counters: Group org.apache.hadoop.mapred.Task$Counter is deprecated. Use org.apache.hadoop.mapreduce.TaskCounter instead
输入行数：0
14/07/15 13:26:48 INFO mapreduce.Job:  map 2% reduce 0%
输入行数：3138474
14/07/15 13:26:51 INFO mapreduce.Job:  map 5% reduce 0%
14/07/15 13:26:54 INFO mapreduce.Job:  map 6% reduce 0%
14/07/15 13:26:55 INFO mapreduce.Job:  map 8% reduce 0%
14/07/15 13:26:57 INFO mapreduce.Job:  map 9% reduce 0%
14/07/15 13:26:58 INFO mapreduce.Job:  map 11% reduce 0%
14/07/15 13:27:00 INFO mapreduce.Job:  map 12% reduce 0%
14/07/15 13:27:01 INFO mapreduce.Job:  map 13% reduce 0%
输入行数：23383595
14/07/15 13:27:05 INFO mapreduce.Job:  map 14% reduce 0%
输入行数：23383595
14/07/15 13:27:23 INFO mapreduce.Job:  map 15% reduce 0%
14/07/15 13:27:27 INFO mapreduce.Job:  map 16% reduce 0%
14/07/15 13:27:28 INFO mapreduce.Job:  map 18% reduce 0%
14/07/15 13:27:30 INFO mapreduce.Job:  map 19% reduce 0%
14/07/15 13:27:31 INFO mapreduce.Job:  map 21% reduce 0%
14/07/15 13:27:34 INFO mapreduce.Job:  map 24% reduce 0%
输入行数：38430301
14/07/15 13:27:37 INFO mapreduce.Job:  map 25% reduce 0%
14/07/15 13:27:40 INFO mapreduce.Job:  map 26% reduce 0%
输入行数：42826322
14/07/15 13:27:57 INFO mapreduce.Job:  map 27% reduce 0%
14/07/15 13:28:00 INFO mapreduce.Job:  map 29% reduce 0%
14/07/15 13:28:02 INFO mapreduce.Job:  map 30% reduce 0%
14/07/15 13:28:03 INFO mapreduce.Job:  map 32% reduce 0%
输入行数：54513531
14/07/15 13:28:05 INFO mapreduce.Job:  map 33% reduce 0%
14/07/15 13:28:06 INFO mapreduce.Job:  map 34% reduce 0%
14/07/15 13:28:08 INFO mapreduce.Job:  map 35% reduce 0%
14/07/15 13:28:09 INFO mapreduce.Job:  map 36% reduce 0%
输入行数：60959081
14/07/15 13:28:22 INFO mapreduce.Job:  map 42% reduce 0%
14/07/15 13:28:30 INFO mapreduce.Job:  map 43% reduce 0%
14/07/15 13:28:31 INFO mapreduce.Job:  map 44% reduce 0%
14/07/15 13:28:34 INFO mapreduce.Job:  map 45% reduce 0%
14/07/15 13:28:35 INFO mapreduce.Job:  map 46% reduce 0%
输入行数：69936159
14/07/15 13:28:37 INFO mapreduce.Job:  map 47% reduce 0%
14/07/15 13:28:38 INFO mapreduce.Job:  map 48% reduce 0%
14/07/15 13:28:41 INFO mapreduce.Job:  map 49% reduce 0%
14/07/15 13:28:44 INFO mapreduce.Job:  map 50% reduce 0%
输入行数：77160461
14/07/15 13:29:01 INFO mapreduce.Job:  map 51% reduce 0%
14/07/15 13:29:04 INFO mapreduce.Job:  map 52% reduce 0%
14/07/15 13:29:05 INFO mapreduce.Job:  map 53% reduce 0%
输入行数：83000373
14/07/15 13:29:07 INFO mapreduce.Job:  map 54% reduce 0%
14/07/15 13:29:09 INFO mapreduce.Job:  map 55% reduce 0%
14/07/15 13:29:10 INFO mapreduce.Job:  map 56% reduce 0%
14/07/15 13:29:13 INFO mapreduce.Job:  map 57% reduce 0%
14/07/15 13:29:16 INFO mapreduce.Job:  map 58% reduce 0%
输入行数：93361766
14/07/15 13:29:32 INFO mapreduce.Job:  map 59% reduce 0%
输入行数：98194696
14/07/15 13:29:35 INFO mapreduce.Job:  map 60% reduce 0%
14/07/15 13:29:37 INFO mapreduce.Job:  map 61% reduce 0%
14/07/15 13:29:38 INFO mapreduce.Job:  map 62% reduce 0%
14/07/15 13:29:40 INFO mapreduce.Job:  map 63% reduce 0%
14/07/15 13:29:41 INFO mapreduce.Job:  map 64% reduce 0%
14/07/15 13:29:44 INFO mapreduce.Job:  map 65% reduce 0%
14/07/15 13:29:48 INFO mapreduce.Job:  map 66% reduce 0%
输入行数：109562184
14/07/15 13:30:04 INFO mapreduce.Job:  map 67% reduce 0%
输入行数：113362818
14/07/15 13:30:06 INFO mapreduce.Job:  map 68% reduce 0%
14/07/15 13:30:08 INFO mapreduce.Job:  map 69% reduce 0%
14/07/15 13:30:10 INFO mapreduce.Job:  map 70% reduce 0%
14/07/15 13:30:12 INFO mapreduce.Job:  map 71% reduce 0%
14/07/15 13:30:15 INFO mapreduce.Job:  map 72% reduce 0%
输入行数：123074119
14/07/15 13:30:32 INFO mapreduce.Job:  map 76% reduce 0%
14/07/15 13:30:33 INFO mapreduce.Job:  map 80% reduce 0%
14/07/15 13:30:34 INFO mapreduce.Job:  map 83% reduce 0%
14/07/15 13:30:35 INFO mapreduce.Job:  map 84% reduce 0%
输入行数：123074119
14/07/15 13:30:37 INFO mapreduce.Job:  map 89% reduce 0%
14/07/15 13:30:38 INFO mapreduce.Job:  map 92% reduce 0%
14/07/15 13:30:39 INFO mapreduce.Job:  map 95% reduce 0%
14/07/15 13:30:40 INFO mapreduce.Job:  map 100% reduce 0%
输入行数：123074119
14/07/15 13:30:53 INFO mapreduce.Job:  map 100% reduce 100%
14/07/15 13:30:53 INFO mapreduce.Job: Job job_1405397597558_0003 completed successfully
14/07/15 13:30:53 INFO mapreduce.Job: Counters: 50
File System Counters
FILE: Number of bytes read=58256119
FILE: Number of bytes written=66039749
FILE: Number of read operations=0
FILE: Number of large read operations=0
FILE: Number of write operations=0
HDFS: Number of bytes read=724520133
HDFS: Number of bytes written=1088895
HDFS: Number of read operations=21
HDFS: Number of large read operations=0
HDFS: Number of write operations=2
Job Counters
Killed map tasks=2
Launched map tasks=8
Launched reduce tasks=1
Data-local map tasks=8
Total time spent by all maps in occupied slots (ms)=1528715
Total time spent by all reduces in occupied slots (ms)=17508
Total time spent by all map tasks (ms)=1528715
Total time spent by all reduce tasks (ms)=17508
Total vcore-seconds taken by all map tasks=1528715
Total vcore-seconds taken by all reduce tasks=17508
Total megabyte-seconds taken by all map tasks=1565404160
Total megabyte-seconds taken by all reduce tasks=17928192
Map-Reduce Framework
Map input records=123074119
Map output records=123074119
Map output bytes=1216795535
Map output materialized bytes=7133406
Input split bytes=594
Combine input records=127374119
Combine output records=4900000
Reduce input groups=100000
Reduce shuffle bytes=7133406
Reduce input records=600000
Reduce output records=100000
Spilled Records=5500000
Shuffled Maps =6
Failed Shuffles=0
Merged Map outputs=6
GC time elapsed (ms)=39761
CPU time spent (ms)=1397060
Physical memory (bytes) snapshot=1797943296
Virtual memory (bytes) snapshot=5082316800
Total committed heap usage (bytes)=1398800384
Shuffle Errors
BAD_ID=0
CONNECTION=0
IO_ERROR=0
WRONG_LENGTH=0
WRONG_MAP=0
WRONG_REDUCE=0
File Input Format Counters
Bytes Read=724519539
File Output Format Counters
Bytes Written=1088895

附录1：WordCount1.java和CounterThread.java的代码

//WordCount1.java代码

package mypackage;

import java.io.IOException;
import java.util.StringTokenizer;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.util.GenericOptionsParser;

public class WordCount1 {
public static class TokenizerMapper extends Mapper<Object, Text, Text, IntWritable>{

private final static IntWritable one = new IntWritable(1);  //建立"int"型变量one,初值为1
private Text word = new Text();                             //建立"string:型变量 word,用于接收传入的单词

public void map(Object key, Text value, Context context
) throws IOException, InterruptedException {
StringTokenizer itr = new StringTokenizer(value.toString());  //将输入的文本按行分段
while (itr.hasMoreTokens()) {
word.set(itr.nextToken());                                  //为word赋值
context.write(word, one);                                   // 将 键-值 对 word one 传入
}
//System.out.println("read lines:"+context.getCounter("org.apache.hadoop.mapred.Task$Counter","MAP_INPUT_RECORDS").getValue());
//System.out.println( "输入行数：" + context.getCounters().findCounter("org.apache.hadoop.mapred.Task$Counter", "MAP_INPUT_RECORDS").getValue() );
//System.out.println( "输入行数：" + context.getCounters().findCounter("", "MAP_INPUT_RECORDS").getValue() );
}
}

public static class IntSumReducer
extends Reducer<Text,IntWritable,Text,IntWritable> {
private IntWritable result = new IntWritable();                 //创建整型变量result

public void reduce(Text key, Iterable<IntWritable> values,
Context context
) throws IOException, InterruptedException {
int sum = 0;                                                 //创建int 型变量sum 初值0
for (IntWritable val : values) {
sum += val.get();                                          //将每个key对应的所有value类间

}
result.set(sum);                                              //sum传入result
context.write(key, result);                                   //将 key-result对传入
}
}

public static void main(String[] args) throws Exception {
Configuration conf = new Configuration();
//String[] newArgs = new String[]{"hdfs://localhost:9000/data/tmpfile","hdfs://localhost:9000/data/wc_output"};
String[] otherArgs = new GenericOptionsParser(conf, args).getRemainingArgs();
if (otherArgs.length != 2) {
System.err.println("Usage: wordcount <in> <out>");
System.exit(2);
}
Job job = new Job(conf, "WordCount1");                  //建立新job
job.setJarByClass(WordCount1.class);
job.setMapperClass(TokenizerMapper.class);              //设置map类
job.setCombinerClass(IntSumReducer.class);              //设置combiner类
job.setReducerClass(IntSumReducer.class);               //设置reducer类
job.setOutputKeyClass(Text.class);                       //输出的key类型
job.setOutputValueClass(IntWritable.class);              //输出的value类型
FileInputFormat.addInputPath(job, new Path(otherArgs[0]));  //输入输出参数(在设置中指定)
FileOutputFormat.setOutputPath(job, new Path(otherArgs[1]));

CounterThread ct = new CounterThread(job);
ct.start();

job.waitForCompletion(true);

System.exit(0);
//System.exit(job.waitForCompletion(true) ? 0 : 1);
}
}

//CounterThread.java的代码

package mypackage;

import java.lang.*;
import java.io.IOException;
import java.util.StringTokenizer;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.JobStatus;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.util.GenericOptionsParser;

public class CounterThread extends Thread{

public CounterThread(Job job) {
_job = job;
}

public void run() {
while(true){
try {
Thread.sleep(1000*5);
} catch (InterruptedException e1) {
// TODO Auto-generated catch block
e1.printStackTrace();
}
try {
if(_job.getStatus().getState() == JobStatus.State.RUNNING)
//continue;
System.out.println( "输入行数：" + _job.getCounters().findCounter("org.apache.hadoop.mapred.Task$Counter", "MAP_INPUT_RECORDS").getValue() );
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (InterruptedException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
}

private Job _job;
}