2010-05-10

Hadoop : Matrix Multiply 用 MapReduce 作矩陣運算

• 測試 ​A MapReduce Algorithm for Matrix Multiplication 所提供之矩陣相乘範例
  • 發現會有 IOException 錯誤訊息，肇因於 /tmp/MatrixMultiply/out/_logs 是目錄而不是檔案

    10/05/10 15:23:23 INFO input.FileInputFormat: Total input paths to process : 1
    10/05/10 15:23:23 INFO mapred.JobClient: Running job: job_201005101012_0016
    10/05/10 15:23:24 INFO mapred.JobClient:  map 0% reduce 0%
    10/05/10 15:23:33 INFO mapred.JobClient:  map 100% reduce 0%
    10/05/10 15:23:45 INFO mapred.JobClient:  map 100% reduce 100%
    10/05/10 15:23:47 INFO mapred.JobClient: Job complete: job_201005101012_0016
    ..........
    Exception in thread "main" java.io.IOException: Cannot open filename /tmp/MatrixMultiply/out/_logs
            at org.apache.hadoop.hdfs.DFSClient$DFSInputStream.openInfo(DFSClient.java:1497)
    

  • [解法] 修改 TestMatrixMultiply.java，並以單機 LocalRunner 執行。( hadoop 0.20.2 預設用 LocalRunner )

    jazz@drbl:~$ wget http://ftp.twaren.net/Unix/Web/apache/hadoop/core/hadoop-0.20.2/hadoop-0.20.2.tar.gz
    jazz@drbl:~$ tar zxvf hadoop-0.20.2.tar.gz
    jazz@drbl:~$ cd hadoop-0.20.2
    jazz@drbl:~/hadoop-0.20.2$ echo "export JAVA_HOME=/usr/lib/jvm/java-6-sun/" >> conf/hadoop-env.sh
    jazz@drbl:~/hadoop-0.20.2$ wget http://trac.nchc.org.tw/grid/raw-attachment/wiki/jazz/10-05-10/matrix.tar.gz
    jazz@drbl:~/hadoop-0.20.2$ tar zxvf matrix.tar.gz
    jazz@drbl:~/hadoop-0.20.2$ cd matrix/
    jazz@drbl:~/hadoop-0.20.2/matrix$ ant
    jazz@drbl:~/hadoop-0.20.2/matrix$ mv matrix.jar ../.
    jazz@drbl:~/hadoop-0.20.2/matrix$ cd ..
    jazz@drbl:~/hadoop-0.20.2$ bin/hadoop jar matrix.jar TestMatrixMultiply
    

    TestMatrixMultiply.java

    @@ -72 +72 @@
                    for (int i = 0; i < rowDim; i++)
                            for (int j = 0; j < colDim; j++)
                                    result[i][j] = 0;
    -               if (fs.isFile(path)) {
    -                       fillMatrix(result, path);
    -               } else {
    -                       FileStatus[] fileStatusArray = fs.listStatus(path);
    -                       for (FileStatus fileStatus : fileStatusArray) {
    -                               fillMatrix(result, fileStatus.getPath());
    -                       }
    +               FileStatus[] fileStatusArray = fs.listStatus(path);
    +               for (FileStatus fileStatus : fileStatusArray) {
    +                 if (fs.isFile(fileStatus.getPath())) {
    +                   fillMatrix(result, fileStatus.getPath());
    +                 }
                    }
                    return result;
            }
    @@ -104 +102 @@
                    int[][] Y = readMatrix(I, J, OUTPUT_DIR_PATH);
                    for (int i = 0; i < I; i++) {
                            for (int j = 0; j < J; j++) {
    +                               System.out.println("......X["+i+"]["+j+"]="+X[i][j]+", Y["+i+"]["+j+"]="+Y[i][j]);
                                    if (X[i][j] != Y[i][j]) {
                                            throw new Exception("Bad answer!");
                                    }
    @@ -283 +282 @@
                    }
            }