alluxio.client.hadoop.DFSIOIntegrationTest.java Source code

Java tutorial

Introduction

Here is the source code for alluxio.client.hadoop.DFSIOIntegrationTest.java

Source

/*
 * The Alluxio Open Foundation licenses this work under the Apache License, version 2.0
 * (the "License"). You may not use this work except in compliance with the License, which is
 * available at www.apache.org/licenses/LICENSE-2.0
 *
 * This software is distributed on an "AS IS" basis, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND,
 * either express or implied, as more fully set forth in the License.
 *
 * See the NOTICE file distributed with this work for information regarding copyright ownership.
 */

package alluxio.client.hadoop;

import alluxio.Constants;
import alluxio.conf.PropertyKey;
import alluxio.conf.ServerConfiguration;
import alluxio.hadoop.FileSystem;
import alluxio.hadoop.HadoopConfigurationUtils;
import alluxio.testutils.BaseIntegrationTest;
import alluxio.testutils.LocalAlluxioClusterResource;

import org.apache.hadoop.fs.FSDataInputStream;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.fs.PositionedReadable;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.SequenceFile;
import org.apache.hadoop.io.SequenceFile.CompressionType;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.io.compress.CompressionCodec;
import org.apache.hadoop.mapred.FileInputFormat;
import org.apache.hadoop.mapred.FileOutputFormat;
import org.apache.hadoop.mapred.JobClient;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapred.Mapper;
import org.apache.hadoop.mapred.OutputCollector;
import org.apache.hadoop.mapred.Reporter;
import org.apache.hadoop.mapred.SequenceFileInputFormat;
import org.apache.hadoop.util.ReflectionUtils;
import org.apache.hadoop.util.StringUtils;
import org.apache.hadoop.util.Tool;
import org.apache.hadoop.util.ToolRunner;
import org.junit.AfterClass;
import org.junit.AssumptionViolatedException;
import org.junit.BeforeClass;
import org.junit.ClassRule;
import org.junit.Test;
import org.junit.rules.TestRule;
import org.junit.runner.Description;
import org.junit.runners.model.Statement;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import java.io.BufferedReader;
import java.io.Closeable;
import java.io.DataInputStream;
import java.io.File;
import java.io.FileOutputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.io.OutputStream;
import java.io.PrintStream;
import java.net.URI;
import java.util.Date;
import java.util.Random;
import java.util.StringTokenizer;

import javax.annotation.Nullable;

/**
 * Distributed i/o benchmark.
 * <p>
 * This test writes into or reads from a specified number of files. Number of bytes to write or read
 * is specified as a parameter to the test. Each file is accessed in a separate map task.
 * <p>
 * The reducer collects the following statistics:
 * <ul>
 * <li>number of tasks completed</li>
 * <li>number of bytes written/read</li>
 * <li>execution time</li>
 * <li>io rate</li>
 * <li>io rate squared</li>
 * </ul>
 *
 * Finally, the following information is appended to a local file
 * <ul>
 * <li>read or write test</li>
 * <li>date and time the test finished</li>
 * <li>number of files</li>
 * <li>total number of bytes processed</li>
 * <li>throughput in mb/sec (total number of bytes / sum of processing times)</li>
 * <li>average i/o rate in mb/sec per file</li>
 * <li>standard deviation of i/o rate</li>
 * </ul>
 */
public class DFSIOIntegrationTest extends BaseIntegrationTest implements Tool {
    /**
     * A rule that is used to enforce supported hadoop client versions before running this test.
     * Value for system property, "alluxio.hadoop.version", is injected by surefire plugin.
     */
    private static class HadoopVersionRule implements TestRule {
        @Override
        public Statement apply(Statement base, Description description) {
            return new Statement() {
                @Override
                public void evaluate() throws Throwable {
                    String hadoopVersion = System.getProperty("alluxio.hadoop.version");
                    if (hadoopVersion != null && (hadoopVersion.startsWith("2.4") || hadoopVersion.startsWith("2.5")
                            || hadoopVersion.startsWith("2.6"))) {
                        throw new AssumptionViolatedException("Hadoop version not supported. Skipping test!");
                    } else {
                        base.evaluate();
                    }
                }
            };
        }
    }

    // Constants for DFSIOIntegrationTest
    private static final Logger LOG = LoggerFactory.getLogger(DFSIOIntegrationTest.class);

    private static final int DEFAULT_BUFFER_SIZE = 4096;
    private static final String BASE_FILE_NAME = "test_io_";
    private static final String DEFAULT_RES_FILE_NAME = "DFSIOIntegrationTest_results.log";
    private static final long MEGA = ByteMultiple.MB.value();
    private static final int DEFAULT_NR_BYTES = 16384;
    private static final int DEFAULT_NR_FILES = 4;
    private static boolean sGenerateReportFile = false;
    private static final String USAGE = "Usage: " + DFSIOIntegrationTest.class.getSimpleName() + " [genericOptions]"
            + " -read [-random | -backward | -skip [-skipSize Size]] |" + " -write | -append | -clean"
            + " [-compression codecClassName]" + " [-nrFiles N]" + " [-size Size[B|KB|MB|GB|TB]]"
            + " [-resFile resultFileName] [-bufferSize Bytes]" + " [-rootDir]";

    private org.apache.hadoop.conf.Configuration mConfig;

    @ClassRule
    public static LocalAlluxioClusterResource sLocalAlluxioClusterResource = new LocalAlluxioClusterResource.Builder()
            .setProperty(PropertyKey.USER_FILE_WRITE_TYPE_DEFAULT, "CACHE_THROUGH").build();
    private static URI sLocalAlluxioClusterUri = null;

    static {
        org.apache.hadoop.conf.Configuration.addDefaultResource("hdfs-default.xml");
        org.apache.hadoop.conf.Configuration.addDefaultResource("hdfs-site.xml");
        org.apache.hadoop.conf.Configuration.addDefaultResource("mapred-default.xml");
        org.apache.hadoop.conf.Configuration.addDefaultResource("mapred-site.xml");
    }

    /**
     * Represents different types of tests.
     */
    private enum TestType {
        TEST_TYPE_READ("read"), TEST_TYPE_WRITE("write"), TEST_TYPE_CLEANUP("cleanup"), TEST_TYPE_APPEND(
                "append"), TEST_TYPE_READ_RANDOM(
                        "random read"), TEST_TYPE_READ_BACKWARD("backward read"), TEST_TYPE_READ_SKIP("skip read");

        private String mType;

        TestType(String t) {
            mType = t;
        }

        @Override
        // String
        public String toString() {
            return mType;
        }
    }

    /**
     * Represents for 5 multiple bytes unit.
     */
    enum ByteMultiple {
        B(1L), KB(0x400L), MB(0x100000L), GB(0x40000000L), TB(0x10000000000L);

        private long mMultiplier;

        ByteMultiple(long mult) {
            mMultiplier = mult;
        }

        long value() {
            return mMultiplier;
        }

        static ByteMultiple parseString(String sMultiple) {
            if (sMultiple == null || sMultiple.isEmpty()) { // MB by default
                return MB;
            }
            String sMU = sMultiple.toUpperCase();
            if (B.name().toUpperCase().endsWith(sMU)) {
                return B;
            }
            if (KB.name().toUpperCase().endsWith(sMU)) {
                return KB;
            }
            if (MB.name().toUpperCase().endsWith(sMU)) {
                return MB;
            }
            if (GB.name().toUpperCase().endsWith(sMU)) {
                return GB;
            }
            if (TB.name().toUpperCase().endsWith(sMU)) {
                return TB;
            }
            throw new IllegalArgumentException("Unsupported ByteMultiple " + sMultiple);
        }
    }

    public DFSIOIntegrationTest() {
        mConfig = new org.apache.hadoop.conf.Configuration();
    }

    private static String getBaseDir(org.apache.hadoop.conf.Configuration conf) {
        return conf.get("test.dfsio.build.data", "/benchmarks/DFSIOIntegrationTest");
    }

    private static Path getControlDir(org.apache.hadoop.conf.Configuration conf) {
        return new Path(getBaseDir(conf), "io_control");
    }

    private static Path getWriteDir(org.apache.hadoop.conf.Configuration conf) {
        return new Path(getBaseDir(conf), "io_write");
    }

    private static Path getReadDir(org.apache.hadoop.conf.Configuration conf) {
        return new Path(getBaseDir(conf), "io_read");
    }

    private static Path getAppendDir(org.apache.hadoop.conf.Configuration conf) {
        return new Path(getBaseDir(conf), "io_append");
    }

    private static Path getRandomReadDir(org.apache.hadoop.conf.Configuration conf) {
        return new Path(getBaseDir(conf), "io_random_read");
    }

    private static Path getDataDir(org.apache.hadoop.conf.Configuration conf) {
        return new Path(getBaseDir(conf), "io_data");
    }

    private static DFSIOIntegrationTest sBench;

    @ClassRule
    public static HadoopVersionRule sHadoopVersionRule = new HadoopVersionRule();

    @BeforeClass
    public static void beforeClass() throws Exception {
        // Init DFSIOIntegrationTest
        sBench = new DFSIOIntegrationTest();
        sBench.getConf().setBoolean("dfs.support.append", true);

        sLocalAlluxioClusterUri = URI.create(sLocalAlluxioClusterResource.get().getMasterURI());
        sBench.getConf().set("fs.defaultFS", sLocalAlluxioClusterUri.toString());
        sBench.getConf().set("fs.default.name", sLocalAlluxioClusterUri.toString());
        sBench.getConf().set("fs." + Constants.SCHEME + ".impl", FileSystem.class.getName());

        org.apache.hadoop.fs.FileSystem fs = org.apache.hadoop.fs.FileSystem.get(sLocalAlluxioClusterUri,
                HadoopConfigurationUtils.mergeAlluxioConfiguration(sBench.getConf(), ServerConfiguration.global()));
        sBench.createControlFile(fs, DEFAULT_NR_BYTES, DEFAULT_NR_FILES);

        /** Check write here, as it is required for other tests */
        writeTest();
    }

    @AfterClass
    public static void afterClass() throws Exception {
        // Clear DFSIOIntegrationTest
        org.apache.hadoop.fs.FileSystem fs = org.apache.hadoop.fs.FileSystem.get(sLocalAlluxioClusterUri,
                HadoopConfigurationUtils.mergeAlluxioConfiguration(sBench.getConf(), ServerConfiguration.global()));
        sBench.cleanup(fs);
    }

    /**
     * Writes into files, then calculates and collects the write test statistics.
     */
    public static void writeTest() throws Exception {
        org.apache.hadoop.fs.FileSystem fs = org.apache.hadoop.fs.FileSystem.get(sLocalAlluxioClusterUri,
                HadoopConfigurationUtils.mergeAlluxioConfiguration(sBench.getConf(), ServerConfiguration.global()));
        long tStart = System.currentTimeMillis();
        sBench.mapperWriteTest(fs);
        long execTime = System.currentTimeMillis() - tStart;
        sBench.analyzeResult(fs, TestType.TEST_TYPE_WRITE, execTime);
    }

    @Test(timeout = 50000)
    public void read() throws Exception {
        org.apache.hadoop.fs.FileSystem fs = org.apache.hadoop.fs.FileSystem.get(sLocalAlluxioClusterUri,
                HadoopConfigurationUtils.mergeAlluxioConfiguration(sBench.getConf(), ServerConfiguration.global()));
        long tStart = System.currentTimeMillis();
        sBench.mapperReadTest(fs);
        long execTime = System.currentTimeMillis() - tStart;
        sBench.analyzeResult(fs, TestType.TEST_TYPE_READ, execTime);
    }

    @Test(timeout = 50000)
    public void readRandom() throws Exception {
        org.apache.hadoop.fs.FileSystem fs = org.apache.hadoop.fs.FileSystem.get(sLocalAlluxioClusterUri,
                HadoopConfigurationUtils.mergeAlluxioConfiguration(sBench.getConf(), ServerConfiguration.global()));
        long tStart = System.currentTimeMillis();
        sBench.getConf().setLong("test.io.skip.size", 0);
        sBench.randomReadTest(fs);
        long execTime = System.currentTimeMillis() - tStart;
        sBench.analyzeResult(fs, TestType.TEST_TYPE_READ_RANDOM, execTime);
    }

    @Test(timeout = 50000)
    public void readBackward() throws Exception {
        org.apache.hadoop.fs.FileSystem fs = org.apache.hadoop.fs.FileSystem.get(sLocalAlluxioClusterUri,
                HadoopConfigurationUtils.mergeAlluxioConfiguration(sBench.getConf(), ServerConfiguration.global()));
        long tStart = System.currentTimeMillis();
        sBench.getConf().setLong("test.io.skip.size", -DEFAULT_BUFFER_SIZE);
        sBench.randomReadTest(fs);
        long execTime = System.currentTimeMillis() - tStart;
        sBench.analyzeResult(fs, TestType.TEST_TYPE_READ_BACKWARD, execTime);
    }

    @Test(timeout = 50000)
    public void readSkip() throws Exception {
        org.apache.hadoop.fs.FileSystem fs = org.apache.hadoop.fs.FileSystem.get(sLocalAlluxioClusterUri,
                HadoopConfigurationUtils.mergeAlluxioConfiguration(sBench.getConf(), ServerConfiguration.global()));
        long tStart = System.currentTimeMillis();
        sBench.getConf().setLong("test.io.skip.size", 1);
        sBench.randomReadTest(fs);
        long execTime = System.currentTimeMillis() - tStart;
        sBench.analyzeResult(fs, TestType.TEST_TYPE_READ_SKIP, execTime);
    }

    @Test(timeout = 50000)
    public void readLargeSkip() throws Exception {
        org.apache.hadoop.fs.FileSystem fs = org.apache.hadoop.fs.FileSystem.get(sLocalAlluxioClusterUri,
                HadoopConfigurationUtils.mergeAlluxioConfiguration(sBench.getConf(), ServerConfiguration.global()));
        long tStart = System.currentTimeMillis();
        sBench.getConf().setLong("test.io.skip.size", 5000);
        sBench.randomReadTest(fs);
        long execTime = System.currentTimeMillis() - tStart;
        sBench.analyzeResult(fs, TestType.TEST_TYPE_READ_SKIP, execTime);
    }

    // TODO(hy): Should active this unit test after ALLUXIO-25 has been solved
    // @Test (timeout = 50000)
    public void append() throws Exception {
        org.apache.hadoop.fs.FileSystem fs = org.apache.hadoop.fs.FileSystem.get(sLocalAlluxioClusterUri,
                HadoopConfigurationUtils.mergeAlluxioConfiguration(sBench.getConf(), ServerConfiguration.global()));
        long tStart = System.currentTimeMillis();
        sBench.mapperAppendTest(fs);
        long execTime = System.currentTimeMillis() - tStart;
        sBench.analyzeResult(fs, TestType.TEST_TYPE_APPEND, execTime);
    }

    @SuppressWarnings("deprecation")
    private void createControlFile(org.apache.hadoop.fs.FileSystem fs, long nrBytes, // in bytes
            int nrFiles) throws IOException {
        LOG.info("creating control file: " + nrBytes + " bytes, " + nrFiles + " files");

        Path controlDir = getControlDir(mConfig);

        if (!fs.exists(controlDir)) {

            fs.delete(controlDir, true);

            for (int i = 0; i < nrFiles; i++) {
                String name = getFileName(i);
                Path controlFile = new Path(controlDir, "in_file_" + name);
                SequenceFile.Writer writer = null;
                try {
                    writer = SequenceFile.createWriter(fs, mConfig, controlFile, Text.class, LongWritable.class,
                            CompressionType.NONE);
                    writer.append(new Text(name), new LongWritable(nrBytes));
                } catch (Exception e) {
                    throw new IOException(e.getLocalizedMessage());
                } finally {
                    if (writer != null) {
                        writer.close();
                    }
                    writer = null;
                }
            }
        }
        LOG.info("created control files for: " + nrFiles + " files");
    }

    private static String getFileName(int fIdx) {
        return BASE_FILE_NAME + Integer.toString(fIdx);
    }

    /**
     * Write/Read mapper base class.
     * <p>
     * Collects the following statistics per task:
     * <ul>
     * <li>number of tasks completed</li>
     * <li>number of bytes written/read</li>
     * <li>execution time</li>
     * <li>i/o rate</li>
     * <li>i/o rate squared</li>
     * </ul>
     */
    private abstract static class IOStatMapper extends AbstractIOMapper<Long> {
        protected CompressionCodec mCompressionCodec;

        IOStatMapper() {
        }

        @Override
        // Mapper
        public void configure(JobConf conf) {
            super.configure(conf);

            // grab compression
            String compression = getConf().get("test.io.compression.class", null);
            Class<? extends CompressionCodec> codec;

            // try to initialize codec
            try {
                codec = (compression == null) ? null
                        : Class.forName(compression).asSubclass(CompressionCodec.class);
            } catch (Exception e) {
                throw new RuntimeException("Compression codec not found: ", e);
            }

            if (codec != null) {
                mCompressionCodec = ReflectionUtils.newInstance(codec, getConf());
            }
        }

        @Override
        // AbstractIOMapper
        void collectStats(OutputCollector<Text, Text> output, String name, long execTime, Long objSize)
                throws IOException {
            long totalSize = objSize;
            float ioRateMbSec = (float) totalSize * 1000 / (execTime * MEGA);
            LOG.info("Number of bytes processed = " + totalSize);
            LOG.info("Exec time = " + execTime);
            LOG.info("IO rate = " + ioRateMbSec);

            output.collect(new Text(AccumulatingReducer.VALUE_TYPE_LONG + "tasks"), new Text(String.valueOf(1)));
            output.collect(new Text(AccumulatingReducer.VALUE_TYPE_LONG + "size"),
                    new Text(String.valueOf(totalSize)));
            output.collect(new Text(AccumulatingReducer.VALUE_TYPE_LONG + "time"),
                    new Text(String.valueOf(execTime)));
            output.collect(new Text(AccumulatingReducer.VALUE_TYPE_FLOAT + "rate"),
                    new Text(String.valueOf(ioRateMbSec * 1000)));
            output.collect(new Text(AccumulatingReducer.VALUE_TYPE_FLOAT + "sqrate"),
                    new Text(String.valueOf(ioRateMbSec * ioRateMbSec * 1000)));
        }
    }

    /**
     * Write mapper class.
     */
    public static class WriteMapper extends IOStatMapper {

        public WriteMapper() {
            for (int i = 0; i < mBufferSize; i++) {
                mBuffer[i] = (byte) ('0' + i % 50);
            }
        }

        @Override
        // AbstractIOMapper
        public Closeable getIOStream(String name) throws IOException {
            // create file
            OutputStream out = mFS.create(new Path(getDataDir(getConf()), name), true, mBufferSize);
            if (mCompressionCodec != null) {
                out = mCompressionCodec.createOutputStream(out);
            }
            LOG.info("out = " + out.getClass().getName());
            return out;
        }

        @Override
        // AbstractIOMapper, totalSize is in bytes
        public Long doIO(Reporter reporter, String name, long totalSize) throws IOException {
            OutputStream out = (OutputStream) this.mStream;
            // write to the file
            long nrRemaining;
            for (nrRemaining = totalSize; nrRemaining > 0; nrRemaining -= mBufferSize) {
                int curSize = (mBufferSize < nrRemaining) ? mBufferSize : (int) nrRemaining;
                out.write(mBuffer, 0, curSize);
                reporter.setStatus("writing " + name + "@" + (totalSize - nrRemaining) + "/" + totalSize
                        + " ::host = " + mHostname);
            }
            return totalSize;
        }
    }

    private void mapperWriteTest(org.apache.hadoop.fs.FileSystem fs) throws IOException {
        Path writeDir = getWriteDir(mConfig);
        fs.delete(getDataDir(mConfig), true);
        fs.delete(writeDir, true);

        runIOTest(WriteMapper.class, writeDir);
    }

    private void runIOTest(Class<? extends Mapper<Text, LongWritable, Text, Text>> mapperClass, Path outputDir)
            throws IOException {
        JobConf job = new JobConf(mConfig, DFSIOIntegrationTest.class);

        FileInputFormat.setInputPaths(job, getControlDir(mConfig));
        job.setInputFormat(SequenceFileInputFormat.class);

        job.setMapperClass(mapperClass);
        job.setReducerClass(AccumulatingReducer.class);

        FileOutputFormat.setOutputPath(job, outputDir);
        job.setOutputKeyClass(Text.class);
        job.setOutputValueClass(Text.class);
        job.setNumReduceTasks(1);
        JobClient.runJob(job);
    }

    /**
     * Append mapper class.
     */
    public static class AppendMapper extends IOStatMapper {

        public AppendMapper() {
            for (int i = 0; i < mBufferSize; i++) {
                mBuffer[i] = (byte) ('0' + i % 50);
            }
        }

        @Override
        // AbstractIOMapper
        public Closeable getIOStream(String name) throws IOException {
            // open file for append
            OutputStream out = mFS.append(new Path(getDataDir(getConf()), name), mBufferSize);
            if (mCompressionCodec != null) {
                out = mCompressionCodec.createOutputStream(out);
            }
            LOG.info("out = " + out.getClass().getName());
            return out;
        }

        @Override
        // AbstractIOMapper, totalSize is in Bytes
        public Long doIO(Reporter reporter, String name, long totalSize) throws IOException {
            OutputStream out = (OutputStream) this.mStream;
            // write to the file
            long nrRemaining;
            for (nrRemaining = totalSize; nrRemaining > 0; nrRemaining -= mBufferSize) {
                int curSize = (mBufferSize < nrRemaining) ? mBufferSize : (int) nrRemaining;
                out.write(mBuffer, 0, curSize);
                reporter.setStatus("writing " + name + "@" + (totalSize - nrRemaining) + "/" + totalSize
                        + " ::host = " + mHostname);
            }
            return totalSize;
        }
    }

    private void mapperAppendTest(org.apache.hadoop.fs.FileSystem fs) throws IOException {
        Path appendDir = getAppendDir(mConfig);
        fs.delete(appendDir, true);
        runIOTest(AppendMapper.class, appendDir);
    }

    /**
     * Read mapper class.
     */
    public static class ReadMapper extends IOStatMapper {

        public ReadMapper() {
        }

        @Override
        // AbstractIOMapper
        public Closeable getIOStream(String name) throws IOException {
            // open file
            InputStream in = mFS.open(new Path(getDataDir(getConf()), name));
            if (mCompressionCodec != null) {
                in = mCompressionCodec.createInputStream(in);
            }
            LOG.info("in = " + in.getClass().getName());
            return in;
        }

        @Override
        // AbstractIOMapper, totalSize in Bytes
        public Long doIO(Reporter reporter, String name, long totalSize) throws IOException {
            InputStream in = (InputStream) this.mStream;
            long actualSize = 0;
            while (actualSize < totalSize) {
                int curSize = in.read(mBuffer, 0, mBufferSize);
                if (curSize < 0) {
                    break;
                }
                actualSize += curSize;
                reporter.setStatus(
                        "reading " + name + "@" + actualSize + "/" + totalSize + " ::host = " + mHostname);
            }
            return actualSize;
        }
    }

    private void mapperReadTest(org.apache.hadoop.fs.FileSystem fs) throws IOException {
        Path readDir = getReadDir(mConfig);
        fs.delete(readDir, true);
        runIOTest(ReadMapper.class, readDir);
    }

    /**
     * Mapper class for random reads. The mapper chooses a position in the file and reads bufferSize
     * bytes starting at the chosen position. It stops after reading the totalSize bytes, specified
     * by size.
     *
     * There are three type of reads. 1) Random read always chooses a random position to read from:
     * skipSize = 0 2) Backward read reads file in reverse order : skipSize &lt; 0 3) Skip-read skips
     * skipSize bytes after every read : skipSize &gt; 0
     */
    public static class RandomReadMapper extends IOStatMapper {
        private Random mRnd;
        private long mFileSize;
        private long mSkipSize;

        @Override
        // Mapper
        public void configure(JobConf conf) {
            super.configure(conf);
            mSkipSize = conf.getLong("test.io.skip.size", 0);
        }

        public RandomReadMapper() {
            mRnd = new Random();
        }

        @Override
        // AbstractIOMapper
        public Closeable getIOStream(String name) throws IOException {
            Path filePath = new Path(getDataDir(getConf()), name);
            mFileSize = mFS.getFileStatus(filePath).getLen();
            InputStream in = mFS.open(filePath);
            if (mCompressionCodec != null) {
                in = new FSDataInputStream(mCompressionCodec.createInputStream(in));
            }
            LOG.info("in = " + in.getClass().getName());
            LOG.info("skipSize = " + mSkipSize);
            return in;
        }

        @Override
        // AbstractIOMapper, totalSize in Bytes
        public Long doIO(Reporter reporter, String name, long totalSize) throws IOException {
            PositionedReadable in = (PositionedReadable) this.mStream;
            long actualSize = 0;
            for (long pos = nextOffset(-1); actualSize < totalSize; pos = nextOffset(pos)) {
                int curSize = in.read(pos, mBuffer, 0, mBufferSize);
                if (curSize < 0) {
                    break;
                }
                actualSize += curSize;
                reporter.setStatus(
                        "reading " + name + "@" + actualSize + "/" + totalSize + " ::host = " + mHostname);
            }
            return actualSize;
        }

        /**
         * Get next offset for reading. If current < 0 then choose initial offset according to the read
         * type.
         *
         * @param current offset
         * @return the next offset for reading
         */
        private long nextOffset(long current) {
            if (mSkipSize == 0) {
                return mRnd.nextInt((int) (mFileSize));
            }
            if (mSkipSize > 0) {
                return (current < 0) ? 0 : (current + mBufferSize + mSkipSize);
            }
            // skipSize < 0
            return (current < 0) ? Math.max(0, mFileSize - mBufferSize) : Math.max(0, current + mSkipSize);
        }
    }

    private void randomReadTest(org.apache.hadoop.fs.FileSystem fs) throws IOException {
        Path readDir = getRandomReadDir(mConfig);
        fs.delete(readDir, true);
        runIOTest(RandomReadMapper.class, readDir);
    }

    // fileSize is in Bytes
    private void sequentialTest(org.apache.hadoop.fs.FileSystem fs, TestType testType, long fileSize, int nrFiles)
            throws IOException {
        IOStatMapper ioer;
        switch (testType) {
        case TEST_TYPE_READ:
            ioer = new ReadMapper();
            break;
        case TEST_TYPE_WRITE:
            ioer = new WriteMapper();
            break;
        case TEST_TYPE_APPEND:
            ioer = new AppendMapper();
            break;
        case TEST_TYPE_READ_RANDOM:
        case TEST_TYPE_READ_BACKWARD:
        case TEST_TYPE_READ_SKIP:
            ioer = new RandomReadMapper();
            break;
        default:
            return;
        }
        for (int i = 0; i < nrFiles; i++) {
            ioer.doIO(Reporter.NULL, BASE_FILE_NAME + Integer.toString(i), fileSize);
        }
        ioer.close();
    }

    /**
     * Runs the integration test for DFS IO.
     *
     * @param args arguments
     */
    public static void main(String[] args) {
        DFSIOIntegrationTest bench = new DFSIOIntegrationTest();
        int res;
        try {
            res = ToolRunner.run(bench, args);
        } catch (Exception e) {
            System.err.print(StringUtils.stringifyException(e));
            res = -2;
        }
        if (res == -1) {
            System.err.print(USAGE);
        }
        System.exit(res);
    }

    @Override
    // Tool
    public int run(String[] args) throws IOException {
        TestType testType = null;
        int bufferSize = DEFAULT_BUFFER_SIZE;
        long nrBytes = MEGA;
        int nrFiles = 1;
        long skipSize = 0;
        String resFileName = DEFAULT_RES_FILE_NAME;
        String compressionClass = null;
        boolean isSequential = false;
        String version = DFSIOIntegrationTest.class.getSimpleName() + ".1.7";
        sGenerateReportFile = true;

        LOG.info(version);
        if (args.length == 0) {
            System.err.println("Missing arguments.");
            return -1;
        }

        for (int i = 0; i < args.length; i++) { // parse command line
            if (args[i].startsWith("-read")) {
                testType = TestType.TEST_TYPE_READ;
            } else if (args[i].equals("-write")) {
                testType = TestType.TEST_TYPE_WRITE;
            } else if (args[i].equals("-append")) {
                testType = TestType.TEST_TYPE_APPEND;
            } else if (args[i].equals("-random")) {
                if (testType != TestType.TEST_TYPE_READ) {
                    return -1;
                }
                testType = TestType.TEST_TYPE_READ_RANDOM;
            } else if (args[i].equals("-backward")) {
                if (testType != TestType.TEST_TYPE_READ) {
                    return -1;
                }
                testType = TestType.TEST_TYPE_READ_BACKWARD;
            } else if (args[i].equals("-skip")) {
                if (testType != TestType.TEST_TYPE_READ) {
                    return -1;
                }
                testType = TestType.TEST_TYPE_READ_SKIP;
            } else if (args[i].equals("-clean")) {
                testType = TestType.TEST_TYPE_CLEANUP;
            } else if (args[i].startsWith("-seq")) {
                isSequential = true;
            } else if (args[i].startsWith("-compression")) {
                compressionClass = args[++i];
            } else if (args[i].equals("-nrFiles")) {
                nrFiles = Integer.parseInt(args[++i]);
            } else if (args[i].equals("-fileSize") || args[i].equals("-size")) {
                nrBytes = parseSize(args[++i]);
            } else if (args[i].equals("-skipSize")) {
                skipSize = parseSize(args[++i]);
            } else if (args[i].equals("-bufferSize")) {
                bufferSize = Integer.parseInt(args[++i]);
            } else if (args[i].equals("-resFile")) {
                resFileName = args[++i];
            } else {
                System.err.println("Illegal argument: " + args[i]);
                return -1;
            }
        }
        if (testType == null) {
            return -1;
        }
        if (testType == TestType.TEST_TYPE_READ_BACKWARD) {
            skipSize = -bufferSize;
        } else if (testType == TestType.TEST_TYPE_READ_SKIP && skipSize == 0) {
            skipSize = bufferSize;
        }

        LOG.info("nrFiles = " + nrFiles);
        LOG.info("nrBytes (MB) = " + toMB(nrBytes));
        LOG.info("bufferSize = " + bufferSize);
        if (skipSize > 0) {
            LOG.info("skipSize = " + skipSize);
        }
        LOG.info("baseDir = " + getBaseDir(mConfig));

        if (compressionClass != null) {
            mConfig.set("test.io.compression.class", compressionClass);
            LOG.info("compressionClass = " + compressionClass);
        }

        mConfig.setInt("test.io.file.buffer.size", bufferSize);
        mConfig.setLong("test.io.skip.size", skipSize);
        mConfig.setBoolean("dfs.support.append", true);
        org.apache.hadoop.fs.FileSystem fs = org.apache.hadoop.fs.FileSystem.get(mConfig);

        if (isSequential) {
            long tStart = System.currentTimeMillis();
            sequentialTest(fs, testType, nrBytes, nrFiles);
            long execTime = System.currentTimeMillis() - tStart;
            String resultLine = "Seq Test exec time sec: " + (float) execTime / 1000;
            LOG.info(resultLine);
            return 0;
        }
        if (testType == TestType.TEST_TYPE_CLEANUP) {
            cleanup(fs);
            return 0;
        }
        createControlFile(fs, nrBytes, nrFiles);
        long tStart = System.currentTimeMillis();
        switch (testType) {
        case TEST_TYPE_WRITE:
            mapperWriteTest(fs);
            break;
        case TEST_TYPE_READ:
            mapperReadTest(fs);
            break;
        case TEST_TYPE_APPEND:
            mapperAppendTest(fs);
            break;
        case TEST_TYPE_READ_RANDOM:
        case TEST_TYPE_READ_BACKWARD:
        case TEST_TYPE_READ_SKIP:
            randomReadTest(fs);
            break;
        default:
        }
        long execTime = System.currentTimeMillis() - tStart;

        analyzeResult(fs, testType, execTime, resFileName);
        return 0;
    }

    @Override
    // Configurable
    public org.apache.hadoop.conf.Configuration getConf() {
        return this.mConfig;
    }

    @Override
    // Configurable
    public void setConf(org.apache.hadoop.conf.Configuration conf) {
        mConfig = conf;
    }

    /**
     * Returns size in bytes.
     *
     * @param arg = {d}[B|KB|MB|GB|TB]
     * @return
     */
    static long parseSize(String arg) {
        String[] args = arg.split("\\D", 2); // get digits
        assert args.length <= 2;
        long nrBytes = Long.parseLong(args[0]);
        String bytesMult = arg.substring(args[0].length()); // get byte multiple
        return nrBytes * ByteMultiple.parseString(bytesMult).value();
    }

    static float toMB(long bytes) {
        return ((float) bytes) / MEGA;
    }

    private void analyzeResult(org.apache.hadoop.fs.FileSystem fs, TestType testType, long execTime,
            String resFileName) throws IOException {
        Path reduceFile = getReduceFilePath(testType);
        long tasks = 0;
        long size = 0;
        long time = 0;
        float rate = 0;
        float sqrate = 0;
        DataInputStream in = null;
        BufferedReader lines = null;
        try {
            in = new DataInputStream(fs.open(reduceFile));
            lines = new BufferedReader(new InputStreamReader(in));
            String line;
            while ((line = lines.readLine()) != null) {
                StringTokenizer tokens = new StringTokenizer(line, " \t\n\r\f%");
                String attr = tokens.nextToken();
                if (attr.endsWith(":tasks")) {
                    tasks = Long.parseLong(tokens.nextToken());
                } else if (attr.endsWith(":size")) {
                    size = Long.parseLong(tokens.nextToken());
                } else if (attr.endsWith(":time")) {
                    time = Long.parseLong(tokens.nextToken());
                } else if (attr.endsWith(":rate")) {
                    rate = Float.parseFloat(tokens.nextToken());
                } else if (attr.endsWith(":sqrate")) {
                    sqrate = Float.parseFloat(tokens.nextToken());
                }
            }
        } finally {
            if (in != null) {
                in.close();
            }
            if (lines != null) {
                lines.close();
            }
        }

        double med = rate / 1000 / tasks;
        double stdDev = Math.sqrt(Math.abs(sqrate / 1000 / tasks - med * med));
        String[] resultLines = { "----- DFSIOIntegrationTest ----- : " + testType,
                "           Date & time: " + new Date(System.currentTimeMillis()),
                "       Number of files: " + tasks, "Total MBytes processed: " + toMB(size),
                "     Throughput mb/sec: " + size * 1000.0 / (time * MEGA), "Average IO rate mb/sec: " + med,
                " IO rate std deviation: " + stdDev, "    Test exec time sec: " + (float) execTime / 1000, "" };

        PrintStream res = null;
        try {
            if (sGenerateReportFile) {
                res = new PrintStream(new FileOutputStream(new File(resFileName), true));
            }
            for (String resultLine : resultLines) {
                LOG.info(resultLine);
                if (sGenerateReportFile) {
                    res.println(resultLine);
                } else {
                    System.out.println(resultLine);
                }
            }
        } finally {
            if (res != null) {
                res.close();
            }
        }
    }

    private void analyzeResult(org.apache.hadoop.fs.FileSystem fs, TestType testType, long execTime)
            throws IOException {
        analyzeResult(fs, testType, execTime, DEFAULT_RES_FILE_NAME);
    }

    @Nullable
    private Path getReduceFilePath(TestType testType) {
        switch (testType) {
        case TEST_TYPE_WRITE:
            return new Path(getWriteDir(mConfig), "part-00000");
        case TEST_TYPE_APPEND:
            return new Path(getAppendDir(mConfig), "part-00000");
        case TEST_TYPE_READ:
            return new Path(getReadDir(mConfig), "part-00000");
        case TEST_TYPE_READ_RANDOM:
        case TEST_TYPE_READ_BACKWARD:
        case TEST_TYPE_READ_SKIP:
            return new Path(getRandomReadDir(mConfig), "part-00000");
        default:
        }
        return null;
    }

    private void cleanup(org.apache.hadoop.fs.FileSystem fs) throws IOException {
        LOG.info("Cleaning up test files");
        fs.delete(new Path(getBaseDir(mConfig)), true);
    }
}