Search in sources :

Example 1 with ShortCircuitCache

use of org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache in project hadoop by apache.

the class TestBlockReaderFactory method testShortCircuitReadFromClientWithoutShm.

/**
   * Test that a client which does not support short-circuit reads using
   * shared memory can talk with a server which supports it.
   */
@Test
public void testShortCircuitReadFromClientWithoutShm() throws Exception {
    TemporarySocketDirectory sockDir = new TemporarySocketDirectory();
    Configuration clientConf = createShortCircuitConf("testShortCircuitReadWithoutShm", sockDir);
    Configuration serverConf = new Configuration(clientConf);
    DFSInputStream.tcpReadsDisabledForTesting = true;
    final MiniDFSCluster cluster = new MiniDFSCluster.Builder(serverConf).numDataNodes(1).build();
    cluster.waitActive();
    clientConf.setInt(DFS_SHORT_CIRCUIT_SHARED_MEMORY_WATCHER_INTERRUPT_CHECK_MS, 0);
    clientConf.set(DFS_CLIENT_CONTEXT, "testShortCircuitReadFromClientWithoutShm_clientContext");
    final DistributedFileSystem fs = (DistributedFileSystem) FileSystem.get(cluster.getURI(0), clientConf);
    final String TEST_FILE = "/test_file";
    final int TEST_FILE_LEN = 4000;
    final int SEED = 0xFADEC;
    DFSTestUtil.createFile(fs, new Path(TEST_FILE), TEST_FILE_LEN, (short) 1, SEED);
    byte[] contents = DFSTestUtil.readFileBuffer(fs, new Path(TEST_FILE));
    byte[] expected = DFSTestUtil.calculateFileContentsFromSeed(SEED, TEST_FILE_LEN);
    Assert.assertTrue(Arrays.equals(contents, expected));
    final ShortCircuitCache cache = fs.getClient().getClientContext().getShortCircuitCache();
    Assert.assertEquals(null, cache.getDfsClientShmManager());
    cluster.shutdown();
    sockDir.close();
}
Also used : Path(org.apache.hadoop.fs.Path) MiniDFSCluster(org.apache.hadoop.hdfs.MiniDFSCluster) Configuration(org.apache.hadoop.conf.Configuration) DatanodeInfoBuilder(org.apache.hadoop.hdfs.protocol.DatanodeInfo.DatanodeInfoBuilder) DistributedFileSystem(org.apache.hadoop.hdfs.DistributedFileSystem) ShortCircuitCache(org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache) TemporarySocketDirectory(org.apache.hadoop.net.unix.TemporarySocketDirectory) Test(org.junit.Test)

Example 2 with ShortCircuitCache

use of org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache in project hadoop by apache.

the class TestBlockReaderFactory method testMultipleWaitersOnShortCircuitCache.

/**
   * Test the case where we have multiple threads waiting on the
   * ShortCircuitCache delivering a certain ShortCircuitReplica.
   *
   * In this case, there should only be one call to
   * createShortCircuitReplicaInfo.  This one replica should be shared
   * by all threads.
   */
@Test(timeout = 60000)
public void testMultipleWaitersOnShortCircuitCache() throws Exception {
    final CountDownLatch latch = new CountDownLatch(1);
    final AtomicBoolean creationIsBlocked = new AtomicBoolean(true);
    final AtomicBoolean testFailed = new AtomicBoolean(false);
    DFSInputStream.tcpReadsDisabledForTesting = true;
    BlockReaderFactory.createShortCircuitReplicaInfoCallback = new ShortCircuitCache.ShortCircuitReplicaCreator() {

        @Override
        public ShortCircuitReplicaInfo createShortCircuitReplicaInfo() {
            Uninterruptibles.awaitUninterruptibly(latch);
            if (!creationIsBlocked.compareAndSet(true, false)) {
                Assert.fail("there were multiple calls to " + "createShortCircuitReplicaInfo.  Only one was expected.");
            }
            return null;
        }
    };
    TemporarySocketDirectory sockDir = new TemporarySocketDirectory();
    Configuration conf = createShortCircuitConf("testMultipleWaitersOnShortCircuitCache", sockDir);
    MiniDFSCluster cluster = new MiniDFSCluster.Builder(conf).numDataNodes(1).build();
    cluster.waitActive();
    final DistributedFileSystem dfs = cluster.getFileSystem();
    final String TEST_FILE = "/test_file";
    final int TEST_FILE_LEN = 4000;
    final int SEED = 0xFADED;
    final int NUM_THREADS = 10;
    DFSTestUtil.createFile(dfs, new Path(TEST_FILE), TEST_FILE_LEN, (short) 1, SEED);
    Runnable readerRunnable = new Runnable() {

        @Override
        public void run() {
            try {
                byte[] contents = DFSTestUtil.readFileBuffer(dfs, new Path(TEST_FILE));
                Assert.assertFalse(creationIsBlocked.get());
                byte[] expected = DFSTestUtil.calculateFileContentsFromSeed(SEED, TEST_FILE_LEN);
                Assert.assertTrue(Arrays.equals(contents, expected));
            } catch (Throwable e) {
                LOG.error("readerRunnable error", e);
                testFailed.set(true);
            }
        }
    };
    Thread[] threads = new Thread[NUM_THREADS];
    for (int i = 0; i < NUM_THREADS; i++) {
        threads[i] = new Thread(readerRunnable);
        threads[i].start();
    }
    Thread.sleep(500);
    latch.countDown();
    for (int i = 0; i < NUM_THREADS; i++) {
        Uninterruptibles.joinUninterruptibly(threads[i]);
    }
    cluster.shutdown();
    sockDir.close();
    Assert.assertFalse(testFailed.get());
}
Also used : Path(org.apache.hadoop.fs.Path) MiniDFSCluster(org.apache.hadoop.hdfs.MiniDFSCluster) Configuration(org.apache.hadoop.conf.Configuration) DatanodeInfoBuilder(org.apache.hadoop.hdfs.protocol.DatanodeInfo.DatanodeInfoBuilder) CountDownLatch(java.util.concurrent.CountDownLatch) ShortCircuitCache(org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache) DistributedFileSystem(org.apache.hadoop.hdfs.DistributedFileSystem) TemporarySocketDirectory(org.apache.hadoop.net.unix.TemporarySocketDirectory) AtomicBoolean(java.util.concurrent.atomic.AtomicBoolean) ShortCircuitReplicaInfo(org.apache.hadoop.hdfs.shortcircuit.ShortCircuitReplicaInfo) Test(org.junit.Test)

Example 3 with ShortCircuitCache

use of org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache in project hadoop by apache.

the class TestBlockReaderFactory method testShortCircuitReadFromServerWithoutShm.

/**
   * Test that a client which supports short-circuit reads using
   * shared memory can fall back to not using shared memory when
   * the server doesn't support it.
   */
@Test
public void testShortCircuitReadFromServerWithoutShm() throws Exception {
    TemporarySocketDirectory sockDir = new TemporarySocketDirectory();
    Configuration clientConf = createShortCircuitConf("testShortCircuitReadFromServerWithoutShm", sockDir);
    Configuration serverConf = new Configuration(clientConf);
    serverConf.setInt(DFS_SHORT_CIRCUIT_SHARED_MEMORY_WATCHER_INTERRUPT_CHECK_MS, 0);
    DFSInputStream.tcpReadsDisabledForTesting = true;
    final MiniDFSCluster cluster = new MiniDFSCluster.Builder(serverConf).numDataNodes(1).build();
    cluster.waitActive();
    clientConf.set(DFS_CLIENT_CONTEXT, "testShortCircuitReadFromServerWithoutShm_clientContext");
    final DistributedFileSystem fs = (DistributedFileSystem) FileSystem.get(cluster.getURI(0), clientConf);
    final String TEST_FILE = "/test_file";
    final int TEST_FILE_LEN = 4000;
    final int SEED = 0xFADEC;
    DFSTestUtil.createFile(fs, new Path(TEST_FILE), TEST_FILE_LEN, (short) 1, SEED);
    byte[] contents = DFSTestUtil.readFileBuffer(fs, new Path(TEST_FILE));
    byte[] expected = DFSTestUtil.calculateFileContentsFromSeed(SEED, TEST_FILE_LEN);
    Assert.assertTrue(Arrays.equals(contents, expected));
    final ShortCircuitCache cache = fs.getClient().getClientContext().getShortCircuitCache();
    final DatanodeInfo datanode = new DatanodeInfoBuilder().setNodeID(cluster.getDataNodes().get(0).getDatanodeId()).build();
    cache.getDfsClientShmManager().visit(new Visitor() {

        @Override
        public void visit(HashMap<DatanodeInfo, PerDatanodeVisitorInfo> info) throws IOException {
            Assert.assertEquals(1, info.size());
            PerDatanodeVisitorInfo vinfo = info.get(datanode);
            Assert.assertTrue(vinfo.disabled);
            Assert.assertEquals(0, vinfo.full.size());
            Assert.assertEquals(0, vinfo.notFull.size());
        }
    });
    cluster.shutdown();
    sockDir.close();
}
Also used : Path(org.apache.hadoop.fs.Path) DatanodeInfo(org.apache.hadoop.hdfs.protocol.DatanodeInfo) MiniDFSCluster(org.apache.hadoop.hdfs.MiniDFSCluster) Configuration(org.apache.hadoop.conf.Configuration) DatanodeInfoBuilder(org.apache.hadoop.hdfs.protocol.DatanodeInfo.DatanodeInfoBuilder) Visitor(org.apache.hadoop.hdfs.shortcircuit.DfsClientShmManager.Visitor) DatanodeInfoBuilder(org.apache.hadoop.hdfs.protocol.DatanodeInfo.DatanodeInfoBuilder) IOException(java.io.IOException) DistributedFileSystem(org.apache.hadoop.hdfs.DistributedFileSystem) ShortCircuitCache(org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache) TemporarySocketDirectory(org.apache.hadoop.net.unix.TemporarySocketDirectory) PerDatanodeVisitorInfo(org.apache.hadoop.hdfs.shortcircuit.DfsClientShmManager.PerDatanodeVisitorInfo) Test(org.junit.Test)

Example 4 with ShortCircuitCache

use of org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache in project hadoop by apache.

the class TestBlockReaderFactory method testShortCircuitCacheShutdown.

/**
   * Test shutting down the ShortCircuitCache while there are things in it.
   */
@Test
public void testShortCircuitCacheShutdown() throws Exception {
    TemporarySocketDirectory sockDir = new TemporarySocketDirectory();
    Configuration conf = createShortCircuitConf("testShortCircuitCacheShutdown", sockDir);
    conf.set(DFS_CLIENT_CONTEXT, "testShortCircuitCacheShutdown");
    Configuration serverConf = new Configuration(conf);
    DFSInputStream.tcpReadsDisabledForTesting = true;
    final MiniDFSCluster cluster = new MiniDFSCluster.Builder(serverConf).numDataNodes(1).build();
    cluster.waitActive();
    final DistributedFileSystem fs = (DistributedFileSystem) FileSystem.get(cluster.getURI(0), conf);
    final String TEST_FILE = "/test_file";
    final int TEST_FILE_LEN = 4000;
    final int SEED = 0xFADEC;
    DFSTestUtil.createFile(fs, new Path(TEST_FILE), TEST_FILE_LEN, (short) 1, SEED);
    byte[] contents = DFSTestUtil.readFileBuffer(fs, new Path(TEST_FILE));
    byte[] expected = DFSTestUtil.calculateFileContentsFromSeed(SEED, TEST_FILE_LEN);
    Assert.assertTrue(Arrays.equals(contents, expected));
    final ShortCircuitCache cache = fs.getClient().getClientContext().getShortCircuitCache();
    cache.close();
    Assert.assertTrue(cache.getDfsClientShmManager().getDomainSocketWatcher().isClosed());
    cluster.shutdown();
    sockDir.close();
}
Also used : Path(org.apache.hadoop.fs.Path) MiniDFSCluster(org.apache.hadoop.hdfs.MiniDFSCluster) Configuration(org.apache.hadoop.conf.Configuration) DatanodeInfoBuilder(org.apache.hadoop.hdfs.protocol.DatanodeInfo.DatanodeInfoBuilder) DistributedFileSystem(org.apache.hadoop.hdfs.DistributedFileSystem) ShortCircuitCache(org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache) TemporarySocketDirectory(org.apache.hadoop.net.unix.TemporarySocketDirectory) Test(org.junit.Test)

Example 5 with ShortCircuitCache

use of org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache in project hadoop by apache.

the class BlockReaderFactory method createShortCircuitReplicaInfo.

/**
   * Fetch a pair of short-circuit block descriptors from a local DataNode.
   *
   * @return    Null if we could not communicate with the datanode,
   *            a new ShortCircuitReplicaInfo object otherwise.
   *            ShortCircuitReplicaInfo objects may contain either an
   *            InvalidToken exception, or a ShortCircuitReplica object ready to
   *            use.
   */
@Override
public ShortCircuitReplicaInfo createShortCircuitReplicaInfo() {
    if (createShortCircuitReplicaInfoCallback != null) {
        ShortCircuitReplicaInfo info = createShortCircuitReplicaInfoCallback.createShortCircuitReplicaInfo();
        if (info != null)
            return info;
    }
    LOG.trace("{}: trying to create ShortCircuitReplicaInfo.", this);
    BlockReaderPeer curPeer;
    while (true) {
        curPeer = nextDomainPeer();
        if (curPeer == null)
            break;
        if (curPeer.fromCache)
            remainingCacheTries--;
        DomainPeer peer = (DomainPeer) curPeer.peer;
        Slot slot = null;
        ShortCircuitCache cache = clientContext.getShortCircuitCache();
        try {
            MutableBoolean usedPeer = new MutableBoolean(false);
            slot = cache.allocShmSlot(datanode, peer, usedPeer, new ExtendedBlockId(block.getBlockId(), block.getBlockPoolId()), clientName);
            if (usedPeer.booleanValue()) {
                LOG.trace("{}: allocShmSlot used up our previous socket {}.  " + "Allocating a new one...", this, peer.getDomainSocket());
                curPeer = nextDomainPeer();
                if (curPeer == null)
                    break;
                peer = (DomainPeer) curPeer.peer;
            }
            ShortCircuitReplicaInfo info = requestFileDescriptors(peer, slot);
            clientContext.getPeerCache().put(datanode, peer);
            return info;
        } catch (IOException e) {
            if (slot != null) {
                cache.freeSlot(slot);
            }
            if (curPeer.fromCache) {
                // Handle an I/O error we got when using a cached socket.
                // These are considered less serious, because the socket may be stale.
                LOG.debug("{}: closing stale domain peer {}", this, peer, e);
                IOUtilsClient.cleanup(LOG, peer);
            } else {
                // Handle an I/O error we got when using a newly created socket.
                // We temporarily disable the domain socket path for a few minutes in
                // this case, to prevent wasting more time on it.
                LOG.warn(this + ": I/O error requesting file descriptors.  " + "Disabling domain socket " + peer.getDomainSocket(), e);
                IOUtilsClient.cleanup(LOG, peer);
                clientContext.getDomainSocketFactory().disableDomainSocketPath(pathInfo.getPath());
                return null;
            }
        }
    }
    return null;
}
Also used : ExtendedBlockId(org.apache.hadoop.hdfs.ExtendedBlockId) MutableBoolean(org.apache.commons.lang.mutable.MutableBoolean) Slot(org.apache.hadoop.hdfs.shortcircuit.ShortCircuitShm.Slot) IOException(java.io.IOException) ShortCircuitReplicaInfo(org.apache.hadoop.hdfs.shortcircuit.ShortCircuitReplicaInfo) ShortCircuitCache(org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache) DomainPeer(org.apache.hadoop.hdfs.net.DomainPeer)

Aggregations

ShortCircuitCache (org.apache.hadoop.hdfs.shortcircuit.ShortCircuitCache)11 MiniDFSCluster (org.apache.hadoop.hdfs.MiniDFSCluster)8 DistributedFileSystem (org.apache.hadoop.hdfs.DistributedFileSystem)7 Test (org.junit.Test)7 Path (org.apache.hadoop.fs.Path)6 Configuration (org.apache.hadoop.conf.Configuration)5 ExtendedBlockId (org.apache.hadoop.hdfs.ExtendedBlockId)5 DatanodeInfoBuilder (org.apache.hadoop.hdfs.protocol.DatanodeInfo.DatanodeInfoBuilder)5 ShortCircuitReplicaInfo (org.apache.hadoop.hdfs.shortcircuit.ShortCircuitReplicaInfo)5 TemporarySocketDirectory (org.apache.hadoop.net.unix.TemporarySocketDirectory)4 IOException (java.io.IOException)3 HdfsConfiguration (org.apache.hadoop.hdfs.HdfsConfiguration)3 ExtendedBlock (org.apache.hadoop.hdfs.protocol.ExtendedBlock)3 ShortCircuitReplica (org.apache.hadoop.hdfs.shortcircuit.ShortCircuitReplica)3 InvalidToken (org.apache.hadoop.security.token.SecretManager.InvalidToken)3 FileInputStream (java.io.FileInputStream)2 ByteBuffer (java.nio.ByteBuffer)2 TimeoutException (java.util.concurrent.TimeoutException)2 AtomicBoolean (java.util.concurrent.atomic.AtomicBoolean)2 MutableBoolean (org.apache.commons.lang.mutable.MutableBoolean)2