Package org.apache.hadoop.hdfs

Source Code of org.apache.hadoop.hdfs.TestFileAppend2

/**
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements.  See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership.  The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License.  You may obtain a copy of the License at
*
*     http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.hadoop.hdfs;

import java.io.IOException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.concurrent.atomic.AtomicReference;

import org.junit.Test;
import static org.junit.Assert.*;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FSDataInputStream;
import org.apache.hadoop.fs.FSDataOutputStream;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.fs.permission.FsPermission;
import org.apache.hadoop.hdfs.server.datanode.DataNode;
import org.apache.hadoop.hdfs.server.datanode.SimulatedFSDataset;
import org.apache.hadoop.hdfs.server.namenode.FSNamesystem;
import org.apache.hadoop.hdfs.server.namenode.LeaseManager;
import org.apache.hadoop.hdfs.server.namenode.NameNode;
import org.apache.hadoop.io.IOUtils;
import org.apache.hadoop.security.AccessControlException;
import org.apache.hadoop.security.UserGroupInformation;

import org.apache.commons.cli.CommandLine;
import org.apache.commons.cli.CommandLineParser;
import org.apache.commons.cli.GnuParser;
import org.apache.commons.cli.HelpFormatter;
import org.apache.commons.cli.OptionBuilder;
import org.apache.commons.cli.Options;
import org.apache.commons.cli.ParseException;
import org.apache.commons.logging.impl.Log4JLogger;
import org.apache.log4j.Level;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;

/**
* This class tests the building blocks that are needed to
* support HDFS appends.
*/
public class TestFileAppend2 {

  {
    ((Log4JLogger)NameNode.stateChangeLog).getLogger().setLevel(Level.ALL);
    ((Log4JLogger)LeaseManager.LOG).getLogger().setLevel(Level.ALL);
    ((Log4JLogger)FSNamesystem.LOG).getLogger().setLevel(Level.ALL);
    ((Log4JLogger)DataNode.LOG).getLogger().setLevel(Level.ALL);
    ((Log4JLogger)DFSClient.LOG).getLogger().setLevel(Level.ALL);
  }

  static Log LOG = LogFactory.getLog(TestFileAppend2.class);

  static final int blockSize = 1024;
  static final int numBlocks = 5;
  static final int fileSize = numBlocks * blockSize + 1;
  boolean simulatedStorage = false;

  private byte[] fileContents = null;

  int numDatanodes = 5;
  int numberOfFiles = 50;
  int numThreads = 10;
  int numAppendsPerThread = 20;
  int artificialBlockReceivedDelay = 50;
  long sleepBetweenSizeChecks = 5000;

  Workload[] workload = null;
  ArrayList<Path> testFiles = new ArrayList<Path>();
  AtomicReference<Throwable> err = new AtomicReference<Throwable>();


  //
  // create a buffer that contains the entire test file data.
  //
  private void initBuffer(int size) {
    long seed = AppendTestUtil.nextLong();
    fileContents = AppendTestUtil.randomBytes(seed, size);
  }

  /*
   * creates a file but does not close it
   */
  private FSDataOutputStream createFile(FileSystem fileSys, Path name, int repl)
    throws IOException {
    FSDataOutputStream stm = fileSys.create(name, true,
                                            fileSys.getConf().getInt("io.file.buffer.size", 4096),
                                            (short)repl, (long)blockSize);
    return stm;
  }

  private void checkFile(FileSystem fs, Path name, int len) throws IOException {
    FSDataInputStream stm = fs.open(name);
    byte[] actual = new byte[len];
    stm.readFully(0, actual);
    checkData(actual, 0, fileContents, "Read 2");
    stm.close();
  }

  private void checkFullFile(FileSystem fs, Path name) throws IOException {
    checkFile(fs, name, fileSize);
  }

  private void checkData(byte[] actual, int from, byte[] expected, String message) {
    for (int idx = 0; idx < actual.length; idx++) {
      assertEquals(message+" byte "+(from+idx)+" differs. expected "+
                   expected[from+idx]+" actual "+actual[idx],
                   expected[from+idx], actual[idx]);
      actual[idx] = 0;
    }
  }


  /**
   * Creates one file, writes a few bytes to it and then closed it.
   * Reopens the same file for appending, write all blocks and then close.
   * Verify that all data exists in file.
   */
  @Test(timeout=200000)
  public void testSimpleAppend() throws IOException {
    final Configuration conf = new Configuration();
    if (simulatedStorage) {
      conf.setBoolean(SimulatedFSDataset.CONFIG_PROPERTY_SIMULATED, true);
    }
    conf.setInt("dfs.datanode.handler.count", 50);
    conf.setBoolean("dfs.support.append", true);
    initBuffer(fileSize);
    MiniDFSCluster cluster = new MiniDFSCluster(conf, 1, true, null);
    FileSystem fs = cluster.getFileSystem();
    try {
      { // test appending to a file.

        // create a new file.
        Path file1 = new Path("/simpleAppend.dat");
        FSDataOutputStream stm = createFile(fs, file1, 1);
        System.out.println("Created file simpleAppend.dat");
 
        // write to file
        int mid = 186;   // io.bytes.per.checksum bytes
        System.out.println("Writing " + mid + " bytes to file " + file1);
        stm.write(fileContents, 0, mid);
        stm.close();
        System.out.println("Wrote and Closed first part of file.");
 
        // write to file
        int mid2 = 607;   // io.bytes.per.checksum bytes
        System.out.println("Writing " + mid + " bytes to file " + file1);
        stm = fs.append(file1);
        stm.write(fileContents, mid, mid2-mid);
        stm.close();
        System.out.println("Wrote and Closed second part of file.");
 
        // write the remainder of the file
        stm = fs.append(file1);

        // ensure getPos is set to reflect existing size of the file
        assertTrue(stm.getPos() > 0);

        System.out.println("Writing " + (fileSize - mid2) + " bytes to file " + file1);
        stm.write(fileContents, mid2, fileSize - mid2);
        System.out.println("Written second part of file");
        stm.close();
        System.out.println("Wrote and Closed second part of file.");
 
        // verify that entire file is good
        checkFullFile(fs, file1);
      }

      { // test appending to an non-existing file.
        FSDataOutputStream out = null;
        try {
          out = fs.append(new Path("/non-existing.dat"));
          fail("Expected to have FileNotFoundException");
        }
        catch(java.io.FileNotFoundException fnfe) {
          System.out.println("Good: got " + fnfe);
          fnfe.printStackTrace(System.out);
        }
        finally {
          IOUtils.closeStream(out);
        }
      }

      { // test append permission.

        //set root to all writable
        Path root = new Path("/");
        fs.setPermission(root, new FsPermission((short)0777));
        fs.close();

        // login as a different user
        final UserGroupInformation superuser = UserGroupInformation.getCurrentUser();
        String username = "testappenduser";
        String group = "testappendgroup";
        assertFalse(superuser.getShortUserName().equals(username));
        assertFalse(Arrays.asList(superuser.getGroupNames()).contains(group));
        UserGroupInformation appenduser =
          UserGroupInformation.createUserForTesting(username, new String[]{group});
       
        fs = DFSTestUtil.getFileSystemAs(appenduser, conf);

        // create a file
        Path dir = new Path(root, getClass().getSimpleName());
        Path foo = new Path(dir, "foo.dat");
        FSDataOutputStream out = null;
        int offset = 0;
        try {
          out = fs.create(foo);
          int len = 10 + AppendTestUtil.nextInt(100);
          out.write(fileContents, offset, len);
          offset += len;
        }
        finally {
          IOUtils.closeStream(out);
        }

        // change dir and foo to minimal permissions.
        fs.setPermission(dir, new FsPermission((short)0100));
        fs.setPermission(foo, new FsPermission((short)0200));

        // try append, should success
        out = null;
        try {
          out = fs.append(foo);
          int len = 10 + AppendTestUtil.nextInt(100);
          out.write(fileContents, offset, len);
          offset += len;
        }
        finally {
          IOUtils.closeStream(out);
        }

        // change dir and foo to all but no write on foo.
        fs.setPermission(foo, new FsPermission((short)0577));
        fs.setPermission(dir, new FsPermission((short)0777));

        // try append, should fail
        out = null;
        try {
          out = fs.append(foo);
          fail("Expected to have AccessControlException");
        }
        catch(AccessControlException ace) {
          System.out.println("Good: got " + ace);
          ace.printStackTrace(System.out);
        }
        finally {
          IOUtils.closeStream(out);
        }
      }
    } catch (IOException e) {
      System.out.println("Exception :" + e);
      throw e;
    } catch (Throwable e) {
      System.out.println("Throwable :" + e);
      e.printStackTrace();
      throw new IOException("Throwable : " + e);
    } finally {
      fs.close();
      cluster.shutdown();
    }
  }

  //
  // an object that does a bunch of appends to files
  //
  class Workload extends Thread {
    private int id;
    private MiniDFSCluster cluster;

    Workload(MiniDFSCluster cluster, int threadIndex) {
      id = threadIndex;
      this.cluster = cluster;
    }

    // create a bunch of files. Write to them and then verify.
    public void run() {
      System.out.println("Workload " + id + " starting... ");
      for (int i = 0; i < numAppendsPerThread && err.get() == null; i++) {
  
        // pick a file at random and remove it from pool
        Path testfile = null;
        synchronized (testFiles) {
          if (testFiles.size() == 0) {
            System.out.println("Completed write to almost all files.");
            return
          }
          int index = AppendTestUtil.nextInt(testFiles.size());
          testfile = testFiles.remove(index);
        }

        long len = 0;
        int sizeToAppend = 0;
        try {
          FileSystem fs = cluster.getFileSystem();

          // add a random number of bytes to file
          len = fs.getFileStatus(testfile).getLen();

          // if file is already full, then pick another file
          if (len >= fileSize) {
            System.out.println("File " + testfile + " is full.");
            continue;
          }
 
          // do small size appends so that we can trigger multiple
          // appends to the same file.
          //
          int left = (int)(fileSize - len)/3;
          if (left <= 0) {
            left = 1;
          }
          sizeToAppend = AppendTestUtil.nextInt(left);

          System.out.println("Workload thread " + id +
                             " appending " + sizeToAppend + " bytes " +
                             " to file " + testfile +
                             " of size " + len);
          FSDataOutputStream stm = fs.append(testfile);
          stm.write(fileContents, (int)len, sizeToAppend);
          stm.close();

          // wait for the file size to be reflected in the namenode metadata
          long startWaitTime = System.currentTimeMillis();
          while (fs.getFileStatus(testfile).getLen() != (len + sizeToAppend)) {
            try {
              System.out.println("Workload thread " + id +
                                 " file " + testfile  +
                                 " size " + fs.getFileStatus(testfile).getLen() +
                                 " expected size " + (len + sizeToAppend) +
                                 " waiting for namenode metadata update.");
              Thread.sleep(sleepBetweenSizeChecks);
              assertTrue("Timed out waiting for len " + (len + sizeToAppend) +
                " in file " + testfile + " (cur len is " +
                fs.getFileStatus(testfile).getLen() + ")",
                System.currentTimeMillis() - startWaitTime < 60000);
            } catch (InterruptedException e) {
            }
          }

          assertTrue("File " + testfile + " size is " +
                     fs.getFileStatus(testfile).getLen() +
                     " but expected " + (len + sizeToAppend),
                    fs.getFileStatus(testfile).getLen() == (len + sizeToAppend));

          checkFile(fs, testfile, (int)(len + sizeToAppend));
        } catch (Throwable e) {
          err.compareAndSet(null, e);
          LOG.error("Workload exception " + id + " testfile " + testfile +
                     " expected size " + (len + sizeToAppend), e);
          return;
        }

        // Add testfile back to the pool of files.
        synchronized (testFiles) {
          testFiles.add(testfile);
        }
      }
    }
  }

  /**
   * Test that appends to files at random offsets.
   */
  @Test
  public void testComplexAppend() throws Throwable {
    initBuffer(fileSize);
    Configuration conf = new Configuration();
    conf.setInt("heartbeat.recheck.interval", 2000);
    conf.setInt("dfs.heartbeat.interval", 2);
    conf.setInt("dfs.replication.pending.timeout.sec", 2);
    conf.setInt("dfs.socket.timeout", 30000);
    conf.setInt("dfs.datanode.socket.write.timeout", 30000);
    conf.setInt("dfs.datanode.handler.count", 50);
    conf.setInt("dfs.datanode.artificialBlockReceivedDelay",
                artificialBlockReceivedDelay);
    conf.setBoolean("dfs.support.append", true);

    MiniDFSCluster cluster = new MiniDFSCluster(conf, numDatanodes,
                                                true, null);
    cluster.waitActive();

    FileSystem fs = cluster.getFileSystem();

    try {
      // create a bunch of test files with random replication factors.
      // Insert them into a linked list.
      //
      for (int i = 0; i < numberOfFiles; i++) {
        short replication = (short)(AppendTestUtil.nextInt(numDatanodes) + 1);
        Path testFile = new Path("/" + i + ".dat");
        FSDataOutputStream stm = createFile(fs, testFile, replication);
        stm.close();
        testFiles.add(testFile);
      }

      // Create threads and make them run workload concurrently.
      workload = new Workload[numThreads];
      for (int i = 0; i < numThreads; i++) {
        workload[i] = new Workload(cluster, i);
        workload[i].setDaemon(true);
        workload[i].start();
      }

      // wait for all transactions to get over
      for (int i = 0; i < numThreads; i++) {
        try {
          System.out.println("Waiting for thread " + i + " to complete...");
          workload[i].join();
          System.out.println("Waiting for thread " + i + " complete.");
        } catch (InterruptedException e) {
          i--;      // retry
        }
      }
    } finally {
      fs.close();
      cluster.shutdown();
    }

    // If any of the worker thread failed in their job, indicate that
    // this test failed.
    //
    if (err.get() != null) {
      throw err.get();
    }
  }
 
  public static final String OPT_NUM_DNS = "numDataNodes";
  public static final String OPT_NUM_FILES = "numFiles";
  public static final String OPT_NUM_THREADS = "numThreads";
  public static final String OPT_NUM_APPENDS = "appendsPerThread";

  @SuppressWarnings("static-access")
  public static void main(String []args) throws Throwable {
    Options options = new Options();
    options.addOption(OptionBuilder
      .withLongOpt(OPT_NUM_DNS).hasArg()
      .withDescription("Number of DNs to start")
      .create());
    options.addOption(OptionBuilder
        .withLongOpt(OPT_NUM_THREADS).hasArg()
        .withDescription("number of threads to append from")
        .create());
    options.addOption(OptionBuilder
        .withLongOpt(OPT_NUM_FILES).hasArg()
        .withDescription("number of files to append to")
        .create());
    options.addOption(OptionBuilder
        .withLongOpt(OPT_NUM_APPENDS).hasArg()
        .withDescription("number of appends per thread")
        .create());
    CommandLineParser parser = new GnuParser();
    CommandLine line;
    try {
      line = parser.parse( options, args );
      if (line.getArgs().length != 0) {
        throw new ParseException("Unexpected options");
      }
    } catch (ParseException pe) {
      HelpFormatter formatter = new HelpFormatter();
      formatter.printHelp("TestFileAppend2", options);
      throw pe;
    }
           
    TestFileAppend2 tfa2 = new TestFileAppend2();
    tfa2.numDatanodes = Integer.parseInt(
        line.getOptionValue(OPT_NUM_DNS, "1"));
    tfa2.numThreads = Integer.parseInt(
        line.getOptionValue(OPT_NUM_THREADS, "30"));
    tfa2.numberOfFiles = Integer.parseInt(
        line.getOptionValue(OPT_NUM_FILES, "1"));
    tfa2.numAppendsPerThread = Integer.parseInt(
        line.getOptionValue(OPT_NUM_APPENDS, "1000"));
   
    // Make workload more aggressive
    tfa2.sleepBetweenSizeChecks = 10;
  
    try {
      tfa2.testComplexAppend();
    } catch (Throwable t) {
      LOG.error("FAILED", t);
      System.exit(1);
    }
    // Something doesn't shut down right about the minicluster
    System.exit(0);
  }
}
TOP

Related Classes of org.apache.hadoop.hdfs.TestFileAppend2

TOP
Copyright © 2018 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.