hbase-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From syuanji...@apache.org
Subject [18/37] hbase git commit: HBASE-15264 Implement a fan out HDFS OutputStream
Date Fri, 26 Feb 2016 23:09:39 GMT
HBASE-15264 Implement a fan out HDFS OutputStream


Project: http://git-wip-us.apache.org/repos/asf/hbase/repo
Commit: http://git-wip-us.apache.org/repos/asf/hbase/commit/6e9d355b
Tree: http://git-wip-us.apache.org/repos/asf/hbase/tree/6e9d355b
Diff: http://git-wip-us.apache.org/repos/asf/hbase/diff/6e9d355b

Branch: refs/heads/hbase-12439
Commit: 6e9d355b12a1e666f4d05be02775a01b6754d063
Parents: a3b4575
Author: zhangduo <zhangduo@apache.org>
Authored: Wed Feb 24 20:47:38 2016 +0800
Committer: zhangduo <zhangduo@apache.org>
Committed: Thu Feb 25 10:07:27 2016 +0800

----------------------------------------------------------------------
 .../util/FanOutOneBlockAsyncDFSOutput.java      | 533 +++++++++++++++
 .../FanOutOneBlockAsyncDFSOutputHelper.java     | 672 +++++++++++++++++++
 ...anOutOneBlockAsyncDFSOutputFlushHandler.java |  61 ++
 .../util/TestFanOutOneBlockAsyncDFSOutput.java  | 190 ++++++
 4 files changed, 1456 insertions(+)
----------------------------------------------------------------------


http://git-wip-us.apache.org/repos/asf/hbase/blob/6e9d355b/hbase-server/src/main/java/org/apache/hadoop/hbase/util/FanOutOneBlockAsyncDFSOutput.java
----------------------------------------------------------------------
diff --git a/hbase-server/src/main/java/org/apache/hadoop/hbase/util/FanOutOneBlockAsyncDFSOutput.java b/hbase-server/src/main/java/org/apache/hadoop/hbase/util/FanOutOneBlockAsyncDFSOutput.java
new file mode 100644
index 0000000..b10f180
--- /dev/null
+++ b/hbase-server/src/main/java/org/apache/hadoop/hbase/util/FanOutOneBlockAsyncDFSOutput.java
@@ -0,0 +1,533 @@
+/**
+ * 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.hbase.util;
+
+import static org.apache.hadoop.hbase.util.FanOutOneBlockAsyncDFSOutputHelper.HEART_BEAT_SEQNO;
+import static org.apache.hadoop.hbase.util.FanOutOneBlockAsyncDFSOutputHelper.completeFile;
+import static org.apache.hadoop.hbase.util.FanOutOneBlockAsyncDFSOutputHelper.endFileLease;
+import static org.apache.hadoop.hbase.util.FanOutOneBlockAsyncDFSOutputHelper.getStatus;
+import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_SOCKET_TIMEOUT_KEY;
+
+import java.io.Closeable;
+import java.io.IOException;
+import java.nio.channels.CompletionHandler;
+import java.util.ArrayDeque;
+import java.util.Collection;
+import java.util.Collections;
+import java.util.Deque;
+import java.util.IdentityHashMap;
+import java.util.List;
+import java.util.Set;
+import java.util.concurrent.Callable;
+import java.util.concurrent.TimeUnit;
+
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.hbase.classification.InterfaceAudience;
+import org.apache.hadoop.hbase.util.FanOutOneBlockAsyncDFSOutputHelper.CancelOnClose;
+import org.apache.hadoop.hdfs.DFSClient;
+import org.apache.hadoop.hdfs.DistributedFileSystem;
+import org.apache.hadoop.hdfs.protocol.ClientProtocol;
+import org.apache.hadoop.hdfs.protocol.DatanodeInfo;
+import org.apache.hadoop.hdfs.protocol.LocatedBlock;
+import org.apache.hadoop.hdfs.protocol.datatransfer.PacketHeader;
+import org.apache.hadoop.hdfs.protocol.datatransfer.PipelineAck;
+import org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.PipelineAckProto;
+import org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.Status;
+import org.apache.hadoop.hdfs.server.common.HdfsServerConstants;
+import org.apache.hadoop.util.DataChecksum;
+
+import com.google.common.base.Supplier;
+
+import io.netty.buffer.ByteBuf;
+import io.netty.buffer.ByteBufAllocator;
+import io.netty.channel.Channel;
+import io.netty.channel.ChannelHandlerContext;
+import io.netty.channel.EventLoop;
+import io.netty.channel.SimpleChannelInboundHandler;
+import io.netty.handler.codec.protobuf.ProtobufDecoder;
+import io.netty.handler.codec.protobuf.ProtobufVarint32FrameDecoder;
+import io.netty.handler.timeout.IdleState;
+import io.netty.handler.timeout.IdleStateEvent;
+import io.netty.handler.timeout.IdleStateHandler;
+import io.netty.util.concurrent.Future;
+import io.netty.util.concurrent.FutureListener;
+import io.netty.util.concurrent.Promise;
+
+/**
+ * An asynchronous HDFS output stream implementation which fans out data to datanode and only
+ * supports writing file with only one block.
+ * <p>
+ * Use the createOutput method in {@link FanOutOneBlockAsyncDFSOutputHelper} to create. The mainly
+ * usage of this class is implementing WAL, so we only expose a little HDFS configurations in the
+ * method. And we place it here under util package because we want to make it independent of WAL
+ * implementation thus easier to move it to HDFS project finally.
+ * <p>
+ * Note that, all connections to datanode will run in the same {@link EventLoop} which means we only
+ * need one thread here. But be careful, we do some blocking operations in {@link #close()} and
+ * {@link #recoverAndClose(CancelableProgressable)} methods, so do not call them inside
+ * {@link EventLoop}. And for {@link #write(byte[])} {@link #write(byte[], int, int)},
+ * {@link #buffered()} and {@link #flush(Object, CompletionHandler, boolean)}, if you call them
+ * outside {@link EventLoop}, there will be an extra context-switch.
+ * <p>
+ * Advantages compare to DFSOutputStream:
+ * <ol>
+ * <li>The fan out mechanism. This will reduce the latency.</li>
+ * <li>The asynchronous WAL could also run in the same EventLoop, we could just call write and flush
+ * inside the EventLoop thread, so generally we only have one thread to do all the things.</li>
+ * <li>Fail-fast when connection to datanode error. The WAL implementation could open new writer
+ * ASAP.</li>
+ * <li>We could benefit from netty's ByteBuf management mechanism.</li>
+ * </ol>
+ */
+@InterfaceAudience.Private
+public class FanOutOneBlockAsyncDFSOutput implements Closeable {
+
+  private final Configuration conf;
+
+  private final FSUtils fsUtils;
+
+  private final DistributedFileSystem dfs;
+
+  private final DFSClient client;
+
+  private final ClientProtocol namenode;
+
+  private final String clientName;
+
+  private final String src;
+
+  private final long fileId;
+
+  private final LocatedBlock locatedBlock;
+
+  private final EventLoop eventLoop;
+
+  private final List<Channel> datanodeList;
+
+  private final DataChecksum summer;
+
+  private final ByteBufAllocator alloc;
+
+  private static final class Callback {
+
+    public final Promise<Void> promise;
+
+    public final long ackedLength;
+
+    public final Set<Channel> unfinishedReplicas;
+
+    public Callback(Promise<Void> promise, long ackedLength, Collection<Channel> replicas) {
+      this.promise = promise;
+      this.ackedLength = ackedLength;
+      if (replicas.isEmpty()) {
+        this.unfinishedReplicas = Collections.emptySet();
+      } else {
+        this.unfinishedReplicas = Collections
+            .newSetFromMap(new IdentityHashMap<Channel, Boolean>(replicas.size()));
+        this.unfinishedReplicas.addAll(replicas);
+      }
+    }
+  }
+
+  private final Deque<Callback> waitingAckQueue = new ArrayDeque<>();
+
+  // this could be different from acked block length because a packet can not start at the middle of
+  // a chunk.
+  private long nextPacketOffsetInBlock = 0L;
+
+  private long nextPacketSeqno = 0L;
+
+  private ByteBuf buf;
+
+  private enum State {
+    STREAMING, CLOSING, BROKEN, CLOSED
+  }
+
+  private State state;
+
+  private void completed(Channel channel) {
+    if (waitingAckQueue.isEmpty()) {
+      return;
+    }
+    for (Callback c : waitingAckQueue) {
+      if (c.unfinishedReplicas.remove(channel)) {
+        if (c.unfinishedReplicas.isEmpty()) {
+          c.promise.trySuccess(null);
+          // since we will remove the Callback entry from waitingAckQueue if its unfinishedReplicas
+          // is empty, so this could only happen at the head of waitingAckQueue, so we just call
+          // removeFirst here.
+          waitingAckQueue.removeFirst();
+          // also wake up flush requests which have the same length.
+          for (Callback cb; (cb = waitingAckQueue.peekFirst()) != null;) {
+            if (cb.ackedLength == c.ackedLength) {
+              cb.promise.trySuccess(null);
+              waitingAckQueue.removeFirst();
+            } else {
+              break;
+            }
+          }
+        }
+        return;
+      }
+    }
+  }
+
+  private void failed(Channel channel, Supplier<Throwable> errorSupplier) {
+    if (state == State.BROKEN || state == State.CLOSED) {
+      return;
+    }
+    if (state == State.CLOSING) {
+      Callback c = waitingAckQueue.peekFirst();
+      if (c == null || !c.unfinishedReplicas.contains(channel)) {
+        // nothing, the endBlock request has already finished.
+        return;
+      }
+    }
+    // disable further write, and fail all pending ack.
+    state = State.BROKEN;
+    Throwable error = errorSupplier.get();
+    for (Callback c : waitingAckQueue) {
+      c.promise.tryFailure(error);
+    }
+    waitingAckQueue.clear();
+    for (Channel ch : datanodeList) {
+      ch.close();
+    }
+  }
+
+  private void setupReceiver(final int timeoutMs) {
+    SimpleChannelInboundHandler<PipelineAckProto> ackHandler = new SimpleChannelInboundHandler<PipelineAckProto>() {
+
+      @Override
+      public boolean isSharable() {
+        return true;
+      }
+
+      @Override
+      protected void channelRead0(final ChannelHandlerContext ctx, PipelineAckProto ack)
+          throws Exception {
+        final Status reply = getStatus(ack);
+        if (reply != Status.SUCCESS) {
+          failed(ctx.channel(), new Supplier<Throwable>() {
+
+            @Override
+            public Throwable get() {
+              return new IOException("Bad response " + reply + " for block "
+                  + locatedBlock.getBlock() + " from datanode " + ctx.channel().remoteAddress());
+            }
+          });
+          return;
+        }
+        if (PipelineAck.isRestartOOBStatus(reply)) {
+          failed(ctx.channel(), new Supplier<Throwable>() {
+
+            @Override
+            public Throwable get() {
+              return new IOException("Restart response " + reply + " for block "
+                  + locatedBlock.getBlock() + " from datanode " + ctx.channel().remoteAddress());
+            }
+          });
+          return;
+        }
+        if (ack.getSeqno() == HEART_BEAT_SEQNO) {
+          return;
+        }
+        completed(ctx.channel());
+      }
+
+      @Override
+      public void channelInactive(final ChannelHandlerContext ctx) throws Exception {
+        failed(ctx.channel(), new Supplier<Throwable>() {
+
+          @Override
+          public Throwable get() {
+            return new IOException("Connection to " + ctx.channel().remoteAddress() + " closed");
+          }
+        });
+      }
+
+      @Override
+      public void exceptionCaught(ChannelHandlerContext ctx, final Throwable cause)
+          throws Exception {
+        failed(ctx.channel(), new Supplier<Throwable>() {
+
+          @Override
+          public Throwable get() {
+            return cause;
+          }
+        });
+      }
+
+      @Override
+      public void userEventTriggered(ChannelHandlerContext ctx, Object evt) throws Exception {
+        if (evt instanceof IdleStateEvent) {
+          IdleStateEvent e = (IdleStateEvent) evt;
+          if (e.state() == IdleState.READER_IDLE) {
+            failed(ctx.channel(), new Supplier<Throwable>() {
+
+              @Override
+              public Throwable get() {
+                return new IOException("Timeout(" + timeoutMs + "ms) waiting for response");
+              }
+            });
+          } else if (e.state() == IdleState.WRITER_IDLE) {
+            PacketHeader heartbeat = new PacketHeader(4, 0, HEART_BEAT_SEQNO, false, 0, false);
+            int len = heartbeat.getSerializedSize();
+            ByteBuf buf = alloc.buffer(len);
+            heartbeat.putInBuffer(buf.nioBuffer(0, len));
+            buf.writerIndex(len);
+            ctx.channel().writeAndFlush(buf);
+          }
+          return;
+        }
+        super.userEventTriggered(ctx, evt);
+      }
+
+    };
+    for (Channel ch : datanodeList) {
+      ch.pipeline().addLast(
+        new IdleStateHandler(timeoutMs, timeoutMs / 2, 0, TimeUnit.MILLISECONDS),
+        new ProtobufVarint32FrameDecoder(),
+        new ProtobufDecoder(PipelineAckProto.getDefaultInstance()), ackHandler);
+      ch.config().setAutoRead(true);
+    }
+  }
+
+  FanOutOneBlockAsyncDFSOutput(Configuration conf, FSUtils fsUtils, DistributedFileSystem dfs,
+      DFSClient client, ClientProtocol namenode, String clientName, String src, long fileId,
+      LocatedBlock locatedBlock, EventLoop eventLoop, List<Channel> datanodeList,
+      DataChecksum summer, ByteBufAllocator alloc) {
+    this.conf = conf;
+    this.fsUtils = fsUtils;
+    this.dfs = dfs;
+    this.client = client;
+    this.namenode = namenode;
+    this.fileId = fileId;
+    this.clientName = clientName;
+    this.src = src;
+    this.locatedBlock = locatedBlock;
+    this.eventLoop = eventLoop;
+    this.datanodeList = datanodeList;
+    this.summer = summer;
+    this.alloc = alloc;
+    this.buf = alloc.directBuffer();
+    this.state = State.STREAMING;
+    setupReceiver(conf.getInt(DFS_CLIENT_SOCKET_TIMEOUT_KEY, HdfsServerConstants.READ_TIMEOUT));
+  }
+
+  /**
+   * Just call write(b, 0, b.length).
+   * @see #write(byte[], int, int)
+   */
+  public void write(byte[] b) {
+    write(b, 0, b.length);
+  }
+
+  /**
+   * Copy the data into the buffer. Note that you need to call
+   * {@link #flush(Object, CompletionHandler, boolean)} to flush the buffer manually.
+   */
+  public void write(final byte[] b, final int off, final int len) {
+    if (eventLoop.inEventLoop()) {
+      buf.ensureWritable(len).writeBytes(b, off, len);
+    } else {
+      eventLoop.submit(new Runnable() {
+
+        @Override
+        public void run() {
+          buf.ensureWritable(len).writeBytes(b, off, len);
+        }
+      }).syncUninterruptibly();
+    }
+  }
+
+  /**
+   * Return the current size of buffered data.
+   */
+  public int buffered() {
+    if (eventLoop.inEventLoop()) {
+      return buf.readableBytes();
+    } else {
+      return eventLoop.submit(new Callable<Integer>() {
+
+        @Override
+        public Integer call() throws Exception {
+          return buf.readableBytes();
+        }
+      }).syncUninterruptibly().getNow().intValue();
+    }
+  }
+
+  public DatanodeInfo[] getPipeline() {
+    return locatedBlock.getLocations();
+  }
+
+  private <A> void flush0(final A attachment, final CompletionHandler<Long, ? super A> handler,
+      boolean syncBlock) {
+    if (state != State.STREAMING) {
+      handler.failed(new IOException("stream already broken"), attachment);
+      return;
+    }
+    int dataLen = buf.readableBytes();
+    final long ackedLength = nextPacketOffsetInBlock + dataLen;
+    if (ackedLength == locatedBlock.getBlock().getNumBytes()) {
+      // no new data, just return
+      handler.completed(locatedBlock.getBlock().getNumBytes(), attachment);
+      return;
+    }
+    Promise<Void> promise = eventLoop.newPromise();
+    promise.addListener(new FutureListener<Void>() {
+
+      @Override
+      public void operationComplete(Future<Void> future) throws Exception {
+        if (future.isSuccess()) {
+          locatedBlock.getBlock().setNumBytes(ackedLength);
+          handler.completed(ackedLength, attachment);
+        } else {
+          handler.failed(future.cause(), attachment);
+        }
+      }
+    });
+    Callback c = waitingAckQueue.peekLast();
+    if (c != null && ackedLength == c.ackedLength) {
+      // just append it to the tail of waiting ack queue,, do not issue new hflush request.
+      waitingAckQueue
+          .addLast(new Callback(promise, ackedLength, Collections.<Channel> emptyList()));
+      return;
+    }
+    int chunkLen = summer.getBytesPerChecksum();
+    int trailingPartialChunkLen = dataLen % chunkLen;
+    int numChecks = dataLen / chunkLen + (trailingPartialChunkLen != 0 ? 1 : 0);
+    int checksumLen = numChecks * summer.getChecksumSize();
+    ByteBuf checksumBuf = alloc.directBuffer(checksumLen);
+    summer.calculateChunkedSums(buf.nioBuffer(), checksumBuf.nioBuffer(0, checksumLen));
+    checksumBuf.writerIndex(checksumLen);
+    PacketHeader header = new PacketHeader(4 + checksumLen + dataLen, nextPacketOffsetInBlock,
+        nextPacketSeqno, false, dataLen, syncBlock);
+    int headerLen = header.getSerializedSize();
+    ByteBuf headerBuf = alloc.buffer(headerLen);
+    header.putInBuffer(headerBuf.nioBuffer(0, headerLen));
+    headerBuf.writerIndex(headerLen);
+
+    waitingAckQueue.addLast(new Callback(promise, ackedLength, datanodeList));
+    for (Channel ch : datanodeList) {
+      ch.write(headerBuf.duplicate().retain());
+      ch.write(checksumBuf.duplicate().retain());
+      ch.writeAndFlush(buf.duplicate().retain());
+    }
+    checksumBuf.release();
+    headerBuf.release();
+    ByteBuf newBuf = alloc.directBuffer().ensureWritable(trailingPartialChunkLen);
+    if (trailingPartialChunkLen != 0) {
+      buf.readerIndex(dataLen - trailingPartialChunkLen).readBytes(newBuf, trailingPartialChunkLen);
+    }
+    buf.release();
+    this.buf = newBuf;
+    nextPacketOffsetInBlock += dataLen - trailingPartialChunkLen;
+    nextPacketSeqno++;
+  }
+
+  /**
+   * Flush the buffer out to datanodes.
+   * @param attachment will be passed to handler when completed.
+   * @param handler will set the acked length as result when completed.
+   * @param syncBlock will call hsync if true, otherwise hflush.
+   */
+  public <A> void flush(final A attachment, final CompletionHandler<Long, ? super A> handler,
+      final boolean syncBlock) {
+    if (eventLoop.inEventLoop()) {
+      flush0(attachment, handler, syncBlock);
+    } else {
+      eventLoop.execute(new Runnable() {
+
+        @Override
+        public void run() {
+          flush0(attachment, handler, syncBlock);
+        }
+      });
+    }
+  }
+
+  private void endBlock(Promise<Void> promise, long size) {
+    if (state != State.STREAMING) {
+      promise.tryFailure(new IOException("stream already broken"));
+      return;
+    }
+    if (!waitingAckQueue.isEmpty()) {
+      promise.tryFailure(new IllegalStateException("should call flush first before calling close"));
+      return;
+    }
+    state = State.CLOSING;
+    PacketHeader header = new PacketHeader(4, size, nextPacketSeqno, true, 0, false);
+    buf.release();
+    buf = null;
+    int headerLen = header.getSerializedSize();
+    ByteBuf headerBuf = alloc.buffer(headerLen);
+    header.putInBuffer(headerBuf.nioBuffer(0, headerLen));
+    headerBuf.writerIndex(headerLen);
+    waitingAckQueue.add(new Callback(promise, size, datanodeList));
+    for (Channel ch : datanodeList) {
+      ch.writeAndFlush(headerBuf.duplicate().retain());
+    }
+    headerBuf.release();
+  }
+
+  /**
+   * The close method when error occurred. Now we just call recoverFileLease.
+   */
+  public void recoverAndClose(CancelableProgressable reporter) throws IOException {
+    assert !eventLoop.inEventLoop();
+    for (Channel ch : datanodeList) {
+      ch.closeFuture().awaitUninterruptibly();
+    }
+    endFileLease(client, src, fileId);
+    fsUtils.recoverFileLease(dfs, new Path(src), conf,
+      reporter == null ? new CancelOnClose(client) : reporter);
+  }
+
+  /**
+   * End the current block and complete file at namenode. You should call
+   * {@link #recoverAndClose(CancelableProgressable)} if this method throws an exception.
+   */
+  @Override
+  public void close() throws IOException {
+    assert !eventLoop.inEventLoop();
+    final Promise<Void> promise = eventLoop.newPromise();
+    eventLoop.execute(new Runnable() {
+
+      @Override
+      public void run() {
+        endBlock(promise, nextPacketOffsetInBlock + buf.readableBytes());
+      }
+    });
+    promise.addListener(new FutureListener<Void>() {
+
+      @Override
+      public void operationComplete(Future<Void> future) throws Exception {
+        for (Channel ch : datanodeList) {
+          ch.close();
+        }
+      }
+    }).syncUninterruptibly();
+    for (Channel ch : datanodeList) {
+      ch.closeFuture().awaitUninterruptibly();
+    }
+    completeFile(client, namenode, src, clientName, locatedBlock.getBlock(), fileId);
+  }
+}

http://git-wip-us.apache.org/repos/asf/hbase/blob/6e9d355b/hbase-server/src/main/java/org/apache/hadoop/hbase/util/FanOutOneBlockAsyncDFSOutputHelper.java
----------------------------------------------------------------------
diff --git a/hbase-server/src/main/java/org/apache/hadoop/hbase/util/FanOutOneBlockAsyncDFSOutputHelper.java b/hbase-server/src/main/java/org/apache/hadoop/hbase/util/FanOutOneBlockAsyncDFSOutputHelper.java
new file mode 100644
index 0000000..d34bbb0
--- /dev/null
+++ b/hbase-server/src/main/java/org/apache/hadoop/hbase/util/FanOutOneBlockAsyncDFSOutputHelper.java
@@ -0,0 +1,672 @@
+/**
+ * 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.hbase.util;
+
+import static io.netty.channel.ChannelOption.CONNECT_TIMEOUT_MILLIS;
+import static org.apache.hadoop.fs.CreateFlag.CREATE;
+import static org.apache.hadoop.fs.CreateFlag.OVERWRITE;
+import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_SOCKET_TIMEOUT_KEY;
+import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_USE_DN_HOSTNAME;
+import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_USE_DN_HOSTNAME_DEFAULT;
+import static org.apache.hadoop.hdfs.protocol.datatransfer.BlockConstructionStage.PIPELINE_SETUP_CREATE;
+
+import java.io.IOException;
+import java.lang.reflect.InvocationTargetException;
+import java.lang.reflect.Method;
+import java.util.ArrayList;
+import java.util.EnumSet;
+import java.util.List;
+import java.util.concurrent.TimeUnit;
+
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.CreateFlag;
+import org.apache.hadoop.fs.FileSystem;
+import org.apache.hadoop.fs.FileSystemLinkResolver;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.fs.UnresolvedLinkException;
+import org.apache.hadoop.fs.permission.FsPermission;
+import org.apache.hadoop.hbase.classification.InterfaceAudience;
+import org.apache.hadoop.hbase.client.ConnectionUtils;
+import org.apache.hadoop.hdfs.DFSClient;
+import org.apache.hadoop.hdfs.DFSOutputStream;
+import org.apache.hadoop.hdfs.DistributedFileSystem;
+import org.apache.hadoop.hdfs.protocol.ClientProtocol;
+import org.apache.hadoop.hdfs.protocol.DatanodeInfo;
+import org.apache.hadoop.hdfs.protocol.ExtendedBlock;
+import org.apache.hadoop.hdfs.protocol.HdfsFileStatus;
+import org.apache.hadoop.hdfs.protocol.LocatedBlock;
+import org.apache.hadoop.hdfs.protocol.datatransfer.BlockConstructionStage;
+import org.apache.hadoop.hdfs.protocol.datatransfer.DataTransferProtoUtil;
+import org.apache.hadoop.hdfs.protocol.datatransfer.DataTransferProtocol;
+import org.apache.hadoop.hdfs.protocol.datatransfer.Op;
+import org.apache.hadoop.hdfs.protocol.datatransfer.PipelineAck;
+import org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.BaseHeaderProto;
+import org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.BlockOpResponseProto;
+import org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.CachingStrategyProto;
+import org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.ChecksumProto;
+import org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.ClientOperationHeaderProto;
+import org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.OpWriteBlockProto;
+import org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.OpWriteBlockProto.Builder;
+import org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.PipelineAckProto;
+import org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.Status;
+import org.apache.hadoop.hdfs.protocol.proto.HdfsProtos.StorageTypeProto;
+import org.apache.hadoop.hdfs.protocolPB.PBHelper;
+import org.apache.hadoop.hdfs.security.token.block.InvalidBlockTokenException;
+import org.apache.hadoop.hdfs.server.common.HdfsServerConstants;
+import org.apache.hadoop.hdfs.server.namenode.LeaseExpiredException;
+import org.apache.hadoop.io.EnumSetWritable;
+import org.apache.hadoop.net.NetUtils;
+import org.apache.hadoop.util.DataChecksum;
+
+import com.google.common.base.Throwables;
+import com.google.common.collect.ImmutableMap;
+import com.google.protobuf.CodedOutputStream;
+
+import io.netty.bootstrap.Bootstrap;
+import io.netty.buffer.ByteBuf;
+import io.netty.buffer.ByteBufAllocator;
+import io.netty.buffer.ByteBufOutputStream;
+import io.netty.buffer.PooledByteBufAllocator;
+import io.netty.channel.Channel;
+import io.netty.channel.ChannelFuture;
+import io.netty.channel.ChannelFutureListener;
+import io.netty.channel.ChannelHandlerContext;
+import io.netty.channel.ChannelInitializer;
+import io.netty.channel.ChannelPipeline;
+import io.netty.channel.EventLoop;
+import io.netty.channel.SimpleChannelInboundHandler;
+import io.netty.channel.socket.nio.NioSocketChannel;
+import io.netty.handler.codec.protobuf.ProtobufDecoder;
+import io.netty.handler.codec.protobuf.ProtobufVarint32FrameDecoder;
+import io.netty.handler.timeout.IdleState;
+import io.netty.handler.timeout.IdleStateEvent;
+import io.netty.handler.timeout.IdleStateHandler;
+import io.netty.util.concurrent.Future;
+import io.netty.util.concurrent.Promise;
+
+/**
+ * Helper class for implementing {@link FanOutOneBlockAsyncDFSOutput}.
+ */
+@InterfaceAudience.Private
+public class FanOutOneBlockAsyncDFSOutputHelper {
+
+  private static final Log LOG = LogFactory.getLog(FanOutOneBlockAsyncDFSOutputHelper.class);
+
+  private FanOutOneBlockAsyncDFSOutputHelper() {
+  }
+
+  // use pooled allocator for performance.
+  private static final ByteBufAllocator ALLOC = PooledByteBufAllocator.DEFAULT;
+
+  // copied from DFSPacket since it is package private.
+  public static final long HEART_BEAT_SEQNO = -1L;
+
+  // helper class for creating DataChecksum object.
+  private static final Method CREATE_CHECKSUM;
+
+  // helper class for getting Status from PipelineAckProto. In hadoop 2.6 or before, there is a
+  // getStatus method, and for hadoop 2.7 or after, the status is retrieved from flag. The flag may
+  // get from proto directly, or combined by the reply field of the proto and a ECN object. See
+  // createPipelineAckStatusGetter for more details.
+  private interface PipelineAckStatusGetter {
+    Status get(PipelineAckProto ack);
+  }
+
+  private static final PipelineAckStatusGetter PIPELINE_ACK_STATUS_GETTER;
+
+  // StorageType enum is added in hadoop 2.4, but it is moved to another package in hadoop 2.6 and
+  // the setter method in OpWriteBlockProto is also added in hadoop 2.6. So we need to skip the
+  // setStorageType call if it is hadoop 2.5 or before. See createStorageTypeSetter for more
+  // details.
+  private interface StorageTypeSetter {
+    OpWriteBlockProto.Builder set(OpWriteBlockProto.Builder builder, Enum<?> storageType);
+  }
+
+  private static final StorageTypeSetter STORAGE_TYPE_SETTER;
+
+  // helper class for calling create method on namenode. There is a supportedVersions parameter for
+  // hadoop 2.6 or after. See createFileCreater for more details.
+  private interface FileCreater {
+    HdfsFileStatus create(ClientProtocol namenode, String src, FsPermission masked,
+        String clientName, EnumSetWritable<CreateFlag> flag, boolean createParent,
+        short replication, long blockSize) throws IOException;
+  }
+
+  private static final FileCreater FILE_CREATER;
+
+  // helper class for add or remove lease from DFSClient. Hadoop 2.4 use src as the Map's key, and
+  // hadoop 2.5 or after use inodeId. See createLeaseManager for more details.
+  private interface LeaseManager {
+
+    void begin(DFSClient client, String src, long inodeId);
+
+    void end(DFSClient client, String src, long inodeId);
+  }
+
+  private static final LeaseManager LEASE_MANAGER;
+
+  // This is used to terminate a recoverFileLease call when FileSystem is already closed.
+  // isClientRunning is not public so we need to use reflection.
+  private interface DFSClientAdaptor {
+    boolean isClientRunning(DFSClient client);
+  }
+
+  private static final DFSClientAdaptor DFS_CLIENT_ADAPTOR;
+
+  private static DFSClientAdaptor createDFSClientAdaptor() {
+    try {
+      final Method method = DFSClient.class.getDeclaredMethod("isClientRunning");
+      method.setAccessible(true);
+      return new DFSClientAdaptor() {
+
+        @Override
+        public boolean isClientRunning(DFSClient client) {
+          try {
+            return (Boolean) method.invoke(client);
+          } catch (IllegalAccessException | InvocationTargetException e) {
+            throw new RuntimeException(e);
+          }
+        }
+      };
+    } catch (NoSuchMethodException e) {
+      throw new Error(e);
+    }
+  }
+
+  private static LeaseManager createLeaseManager() {
+    try {
+      final Method beginFileLeaseMethod = DFSClient.class.getDeclaredMethod("beginFileLease",
+        long.class, DFSOutputStream.class);
+      beginFileLeaseMethod.setAccessible(true);
+      final Method endFileLeaseMethod = DFSClient.class.getDeclaredMethod("endFileLease",
+        long.class);
+      endFileLeaseMethod.setAccessible(true);
+      return new LeaseManager() {
+
+        @Override
+        public void begin(DFSClient client, String src, long inodeId) {
+          try {
+            beginFileLeaseMethod.invoke(client, inodeId, null);
+          } catch (IllegalAccessException | InvocationTargetException e) {
+            throw new RuntimeException(e);
+          }
+        }
+
+        @Override
+        public void end(DFSClient client, String src, long inodeId) {
+          try {
+            endFileLeaseMethod.invoke(client, inodeId);
+          } catch (IllegalAccessException | InvocationTargetException e) {
+            throw new RuntimeException(e);
+          }
+        }
+      };
+    } catch (NoSuchMethodException e) {
+      LOG.warn("No inodeId related lease methods found, should be hadoop 2.4-", e);
+    }
+    try {
+      final Method beginFileLeaseMethod = DFSClient.class.getDeclaredMethod("beginFileLease",
+        String.class, DFSOutputStream.class);
+      beginFileLeaseMethod.setAccessible(true);
+      final Method endFileLeaseMethod = DFSClient.class.getDeclaredMethod("endFileLease",
+        String.class);
+      endFileLeaseMethod.setAccessible(true);
+      return new LeaseManager() {
+
+        @Override
+        public void begin(DFSClient client, String src, long inodeId) {
+          try {
+            beginFileLeaseMethod.invoke(client, src, null);
+          } catch (IllegalAccessException | InvocationTargetException e) {
+            throw new RuntimeException(e);
+          }
+        }
+
+        @Override
+        public void end(DFSClient client, String src, long inodeId) {
+          try {
+            endFileLeaseMethod.invoke(client, src);
+          } catch (IllegalAccessException | InvocationTargetException e) {
+            throw new RuntimeException(e);
+          }
+        }
+      };
+    } catch (NoSuchMethodException e) {
+      throw new Error(e);
+    }
+  }
+
+  private static PipelineAckStatusGetter createPipelineAckStatusGetter() {
+    try {
+      final Method getFlagListMethod = PipelineAckProto.class.getMethod("getFlagList");
+      @SuppressWarnings("rawtypes")
+      Class<? extends Enum> ecnClass;
+      try {
+        ecnClass = Class.forName("org.apache.hadoop.hdfs.protocol.datatransfer.PipelineAck$ECN")
+            .asSubclass(Enum.class);
+      } catch (ClassNotFoundException e) {
+        throw new Error(e);
+      }
+      @SuppressWarnings("unchecked")
+      final Enum<?> disabledECN = Enum.valueOf(ecnClass, "DISABLED");
+      final Method getReplyMethod = PipelineAckProto.class.getMethod("getReply", int.class);
+      final Method combineHeaderMethod = PipelineAck.class.getMethod("combineHeader", ecnClass,
+        Status.class);
+      final Method getStatusFromHeaderMethod = PipelineAck.class.getMethod("getStatusFromHeader",
+        int.class);
+      return new PipelineAckStatusGetter() {
+
+        @Override
+        public Status get(PipelineAckProto ack) {
+          try {
+            @SuppressWarnings("unchecked")
+            List<Integer> flagList = (List<Integer>) getFlagListMethod.invoke(ack);
+            Integer headerFlag;
+            if (flagList.isEmpty()) {
+              Status reply = (Status) getReplyMethod.invoke(ack, 0);
+              headerFlag = (Integer) combineHeaderMethod.invoke(null, disabledECN, reply);
+            } else {
+              headerFlag = flagList.get(0);
+            }
+            return (Status) getStatusFromHeaderMethod.invoke(null, headerFlag);
+          } catch (IllegalAccessException | InvocationTargetException e) {
+            throw new RuntimeException(e);
+          }
+        }
+      };
+    } catch (NoSuchMethodException e) {
+      LOG.warn("Can not get expected methods, should be hadoop 2.6-", e);
+    }
+    try {
+      final Method getStatusMethod = PipelineAckProto.class.getMethod("getStatus", int.class);
+      return new PipelineAckStatusGetter() {
+
+        @Override
+        public Status get(PipelineAckProto ack) {
+          try {
+            return (Status) getStatusMethod.invoke(ack, 0);
+          } catch (IllegalAccessException | InvocationTargetException e) {
+            throw new RuntimeException(e);
+          }
+        }
+      };
+    } catch (NoSuchMethodException e) {
+      throw new Error(e);
+    }
+  }
+
+  private static StorageTypeSetter createStorageTypeSetter() {
+    final Method setStorageTypeMethod;
+    try {
+      setStorageTypeMethod = OpWriteBlockProto.Builder.class.getMethod("setStorageType",
+        StorageTypeProto.class);
+    } catch (NoSuchMethodException e) {
+      LOG.warn("noSetStorageType method found, should be hadoop 2.5-", e);
+      return new StorageTypeSetter() {
+
+        @Override
+        public Builder set(Builder builder, Enum<?> storageType) {
+          return builder;
+        }
+      };
+    }
+    ImmutableMap.Builder<String, StorageTypeProto> builder = ImmutableMap.builder();
+    for (StorageTypeProto storageTypeProto : StorageTypeProto.values()) {
+      builder.put(storageTypeProto.name(), storageTypeProto);
+    }
+    final ImmutableMap<String, StorageTypeProto> name2ProtoEnum = builder.build();
+    return new StorageTypeSetter() {
+
+      @Override
+      public Builder set(Builder builder, Enum<?> storageType) {
+        Object protoEnum = name2ProtoEnum.get(storageType.name());
+        try {
+          setStorageTypeMethod.invoke(builder, protoEnum);
+        } catch (IllegalAccessException | IllegalArgumentException | InvocationTargetException e) {
+          throw new RuntimeException(e);
+        }
+        return builder;
+      }
+    };
+  }
+
+  private static FileCreater createFileCreater() {
+    for (Method method : ClientProtocol.class.getMethods()) {
+      if (method.getName().equals("create")) {
+        final Method createMethod = method;
+        Class<?>[] paramTypes = createMethod.getParameterTypes();
+        if (paramTypes[paramTypes.length - 1] == long.class) {
+          return new FileCreater() {
+
+            @Override
+            public HdfsFileStatus create(ClientProtocol namenode, String src, FsPermission masked,
+                String clientName, EnumSetWritable<CreateFlag> flag, boolean createParent,
+                short replication, long blockSize) throws IOException {
+              try {
+                return (HdfsFileStatus) createMethod.invoke(namenode, src, masked, clientName, flag,
+                  createParent, replication, blockSize);
+              } catch (IllegalAccessException e) {
+                throw new RuntimeException(e);
+              } catch (InvocationTargetException e) {
+                Throwables.propagateIfPossible(e.getTargetException(), IOException.class);
+                throw new RuntimeException(e);
+              }
+            }
+          };
+        } else {
+          try {
+            Class<?> cryptoProtocolVersionClass = Class
+                .forName("org.apache.hadoop.crypto.CryptoProtocolVersion");
+            Method supportedMethod = cryptoProtocolVersionClass.getMethod("supported");
+            final Object supported = supportedMethod.invoke(null);
+            return new FileCreater() {
+
+              @Override
+              public HdfsFileStatus create(ClientProtocol namenode, String src, FsPermission masked,
+                  String clientName, EnumSetWritable<CreateFlag> flag, boolean createParent,
+                  short replication, long blockSize) throws IOException {
+                try {
+                  return (HdfsFileStatus) createMethod.invoke(namenode, src, masked, clientName,
+                    flag, createParent, replication, blockSize, supported);
+                } catch (IllegalAccessException e) {
+                  throw new RuntimeException(e);
+                } catch (InvocationTargetException e) {
+                  Throwables.propagateIfPossible(e.getTargetException(), IOException.class);
+                  throw new RuntimeException(e);
+                }
+              }
+            };
+          } catch (ClassNotFoundException | NoSuchMethodException | IllegalAccessException
+              | InvocationTargetException e) {
+            throw new Error(e);
+          }
+        }
+      }
+    }
+    throw new Error("No create method found for " + ClientProtocol.class.getName());
+  }
+
+  // cancel the processing if DFSClient is already closed.
+  static final class CancelOnClose implements CancelableProgressable {
+
+    private final DFSClient client;
+
+    public CancelOnClose(DFSClient client) {
+      this.client = client;
+    }
+
+    @Override
+    public boolean progress() {
+      return DFS_CLIENT_ADAPTOR.isClientRunning(client);
+    }
+
+  }
+
+  static {
+    try {
+      CREATE_CHECKSUM = DFSClient.Conf.class.getDeclaredMethod("createChecksum");
+      CREATE_CHECKSUM.setAccessible(true);
+    } catch (NoSuchMethodException e) {
+      throw new Error(e);
+    }
+
+    PIPELINE_ACK_STATUS_GETTER = createPipelineAckStatusGetter();
+    STORAGE_TYPE_SETTER = createStorageTypeSetter();
+    FILE_CREATER = createFileCreater();
+    LEASE_MANAGER = createLeaseManager();
+    DFS_CLIENT_ADAPTOR = createDFSClientAdaptor();
+  }
+
+  static void beginFileLease(DFSClient client, String src, long inodeId) {
+    LEASE_MANAGER.begin(client, src, inodeId);
+  }
+
+  static void endFileLease(DFSClient client, String src, long inodeId) {
+    LEASE_MANAGER.end(client, src, inodeId);
+  }
+
+  static DataChecksum createChecksum(DFSClient client) {
+    try {
+      return (DataChecksum) CREATE_CHECKSUM.invoke(client.getConf());
+    } catch (IllegalAccessException | InvocationTargetException e) {
+      throw new RuntimeException(e);
+    }
+  }
+
+  static Status getStatus(PipelineAckProto ack) {
+    return PIPELINE_ACK_STATUS_GETTER.get(ack);
+  }
+
+  private static void processWriteBlockResponse(Channel channel, final DatanodeInfo dnInfo,
+      final Promise<Channel> promise, final int timeoutMs) {
+    channel.pipeline().addLast(new IdleStateHandler(timeoutMs, 0, 0, TimeUnit.MILLISECONDS),
+      new ProtobufVarint32FrameDecoder(),
+      new ProtobufDecoder(BlockOpResponseProto.getDefaultInstance()),
+      new SimpleChannelInboundHandler<BlockOpResponseProto>() {
+
+        @Override
+        protected void channelRead0(ChannelHandlerContext ctx, BlockOpResponseProto resp)
+            throws Exception {
+          Status pipelineStatus = resp.getStatus();
+          if (PipelineAck.isRestartOOBStatus(pipelineStatus)) {
+            throw new IOException("datanode " + dnInfo + " is restarting");
+          }
+          String logInfo = "ack with firstBadLink as " + resp.getFirstBadLink();
+          if (resp.getStatus() != Status.SUCCESS) {
+            if (resp.getStatus() == Status.ERROR_ACCESS_TOKEN) {
+              throw new InvalidBlockTokenException("Got access token error" + ", status message "
+                  + resp.getMessage() + ", " + logInfo);
+            } else {
+              throw new IOException("Got error" + ", status=" + resp.getStatus().name()
+                  + ", status message " + resp.getMessage() + ", " + logInfo);
+            }
+          }
+          // success
+          ChannelPipeline p = ctx.pipeline();
+          while (p.first() != null) {
+            p.removeFirst();
+          }
+          // Disable auto read here. Enable it after we setup the streaming pipeline in
+          // FanOutOneBLockAsyncDFSOutput.
+          ctx.channel().config().setAutoRead(false);
+          promise.trySuccess(ctx.channel());
+        }
+
+        @Override
+        public void channelInactive(ChannelHandlerContext ctx) throws Exception {
+          promise.tryFailure(new IOException("connection to " + dnInfo + " is closed"));
+        }
+
+        @Override
+        public void userEventTriggered(ChannelHandlerContext ctx, Object evt) throws Exception {
+          if (evt instanceof IdleStateEvent
+              && ((IdleStateEvent) evt).state() == IdleState.READER_IDLE) {
+            promise
+                .tryFailure(new IOException("Timeout(" + timeoutMs + "ms) waiting for response"));
+          } else {
+            super.userEventTriggered(ctx, evt);
+          }
+        }
+
+        @Override
+        public void exceptionCaught(ChannelHandlerContext ctx, Throwable cause) throws Exception {
+          promise.tryFailure(cause);
+        }
+      });
+  }
+
+  private static void requestWriteBlock(Channel channel, Enum<?> storageType,
+      OpWriteBlockProto.Builder writeBlockProtoBuilder) throws IOException {
+    // TODO: SASL negotiation. should be done using a netty Handler.
+    OpWriteBlockProto proto = STORAGE_TYPE_SETTER.set(writeBlockProtoBuilder, storageType).build();
+    int protoLen = proto.getSerializedSize();
+    ByteBuf buffer = channel.alloc()
+        .buffer(3 + CodedOutputStream.computeRawVarint32Size(protoLen) + protoLen);
+    buffer.writeShort(DataTransferProtocol.DATA_TRANSFER_VERSION);
+    buffer.writeByte(Op.WRITE_BLOCK.code);
+    proto.writeDelimitedTo(new ByteBufOutputStream(buffer));
+    channel.writeAndFlush(buffer);
+  }
+
+  private static List<Future<Channel>> connectToDataNodes(Configuration conf, String clientName,
+      LocatedBlock locatedBlock, long maxBytesRcvd, long latestGS, BlockConstructionStage stage,
+      DataChecksum summer, EventLoop eventLoop) {
+    Enum<?>[] storageTypes = locatedBlock.getStorageTypes();
+    DatanodeInfo[] datanodeInfos = locatedBlock.getLocations();
+    boolean connectToDnViaHostname = conf.getBoolean(DFS_CLIENT_USE_DN_HOSTNAME,
+      DFS_CLIENT_USE_DN_HOSTNAME_DEFAULT);
+    final int timeoutMs = conf.getInt(DFS_CLIENT_SOCKET_TIMEOUT_KEY,
+      HdfsServerConstants.READ_TIMEOUT);
+    ExtendedBlock blockCopy = new ExtendedBlock(locatedBlock.getBlock());
+    blockCopy.setNumBytes(locatedBlock.getBlockSize());
+    ClientOperationHeaderProto header = ClientOperationHeaderProto.newBuilder()
+        .setBaseHeader(BaseHeaderProto.newBuilder().setBlock(PBHelper.convert(blockCopy))
+            .setToken(PBHelper.convert(locatedBlock.getBlockToken())))
+        .setClientName(clientName).build();
+    ChecksumProto checksumProto = DataTransferProtoUtil.toProto(summer);
+    final OpWriteBlockProto.Builder writeBlockProtoBuilder = OpWriteBlockProto.newBuilder()
+        .setHeader(header).setStage(OpWriteBlockProto.BlockConstructionStage.valueOf(stage.name()))
+        .setPipelineSize(1).setMinBytesRcvd(locatedBlock.getBlock().getNumBytes())
+        .setMaxBytesRcvd(maxBytesRcvd).setLatestGenerationStamp(latestGS)
+        .setRequestedChecksum(checksumProto)
+        .setCachingStrategy(CachingStrategyProto.newBuilder().setDropBehind(true).build());
+    List<Future<Channel>> futureList = new ArrayList<>(datanodeInfos.length);
+    for (int i = 0; i < datanodeInfos.length; i++) {
+      final DatanodeInfo dnInfo = datanodeInfos[i];
+      // Use Enum here because StoregType is moved to another package in hadoop 2.6. Use StorageType
+      // will cause compilation error for hadoop 2.5 or before.
+      final Enum<?> storageType = storageTypes[i];
+      final Promise<Channel> promise = eventLoop.newPromise();
+      futureList.add(promise);
+      String dnAddr = dnInfo.getXferAddr(connectToDnViaHostname);
+      new Bootstrap().group(eventLoop).channel(NioSocketChannel.class)
+          .option(CONNECT_TIMEOUT_MILLIS, timeoutMs).handler(new ChannelInitializer<Channel>() {
+
+            @Override
+            protected void initChannel(Channel ch) throws Exception {
+              processWriteBlockResponse(ch, dnInfo, promise, timeoutMs);
+            }
+          }).connect(NetUtils.createSocketAddr(dnAddr)).addListener(new ChannelFutureListener() {
+
+            @Override
+            public void operationComplete(ChannelFuture future) throws Exception {
+              if (future.isSuccess()) {
+                requestWriteBlock(future.channel(), storageType, writeBlockProtoBuilder);
+              } else {
+                promise.tryFailure(future.cause());
+              }
+            }
+          });
+    }
+    return futureList;
+  }
+
+  private static FanOutOneBlockAsyncDFSOutput createOutput(DistributedFileSystem dfs, String src,
+      boolean overwrite, boolean createParent, short replication, long blockSize,
+      EventLoop eventLoop) throws IOException {
+    Configuration conf = dfs.getConf();
+    FSUtils fsUtils = FSUtils.getInstance(dfs, conf);
+    DFSClient client = dfs.getClient();
+    String clientName = client.getClientName();
+    ClientProtocol namenode = client.getNamenode();
+    HdfsFileStatus stat = FILE_CREATER.create(namenode, src,
+      FsPermission.getFileDefault().applyUMask(FsPermission.getUMask(conf)), clientName,
+      new EnumSetWritable<CreateFlag>(
+          overwrite ? EnumSet.of(CREATE, OVERWRITE) : EnumSet.of(CREATE)),
+      createParent, replication, blockSize);
+    beginFileLease(client, src, stat.getFileId());
+    boolean succ = false;
+    LocatedBlock locatedBlock = null;
+    List<Channel> datanodeList = new ArrayList<>();
+    try {
+      DataChecksum summer = createChecksum(client);
+      locatedBlock = namenode.addBlock(src, client.getClientName(), null, null, stat.getFileId(),
+        null);
+      for (Future<Channel> future : connectToDataNodes(conf, clientName, locatedBlock, 0L, 0L,
+        PIPELINE_SETUP_CREATE, summer, eventLoop)) {
+        // fail the creation if there are connection failures since we are fail-fast. The upper
+        // layer should retry itself if needed.
+        datanodeList.add(future.syncUninterruptibly().getNow());
+      }
+      succ = true;
+      return new FanOutOneBlockAsyncDFSOutput(conf, fsUtils, dfs, client, namenode, clientName, src,
+          stat.getFileId(), locatedBlock, eventLoop, datanodeList, summer, ALLOC);
+    } finally {
+      if (!succ) {
+        for (Channel c : datanodeList) {
+          c.close();
+        }
+        endFileLease(client, src, stat.getFileId());
+        fsUtils.recoverFileLease(dfs, new Path(src), conf, new CancelOnClose(client));
+      }
+    }
+  }
+
+  /**
+   * Create a {@link FanOutOneBlockAsyncDFSOutput}. The method maybe blocked so do not call it
+   * inside {@link EventLoop}.
+   * @param eventLoop all connections to datanode will use the same event loop.
+   */
+  public static FanOutOneBlockAsyncDFSOutput createOutput(final DistributedFileSystem dfs, Path f,
+      final boolean overwrite, final boolean createParent, final short replication,
+      final long blockSize, final EventLoop eventLoop) throws IOException {
+    return new FileSystemLinkResolver<FanOutOneBlockAsyncDFSOutput>() {
+
+      @Override
+      public FanOutOneBlockAsyncDFSOutput doCall(Path p)
+          throws IOException, UnresolvedLinkException {
+        return createOutput(dfs, p.toUri().getPath(), overwrite, createParent, replication,
+          blockSize, eventLoop);
+      }
+
+      @Override
+      public FanOutOneBlockAsyncDFSOutput next(FileSystem fs, Path p) throws IOException {
+        throw new UnsupportedOperationException();
+      }
+    }.resolve(dfs, f);
+  }
+
+  static void completeFile(DFSClient client, ClientProtocol namenode, String src, String clientName,
+      ExtendedBlock block, long fileId) {
+    for (int retry = 0;; retry++) {
+      try {
+        if (namenode.complete(src, clientName, block, fileId)) {
+          endFileLease(client, src, fileId);
+          return;
+        } else {
+          LOG.warn("complete file " + src + " not finished, retry = " + retry);
+        }
+      } catch (LeaseExpiredException e) {
+        LOG.warn("lease for file " + src + " is expired, give up", e);
+        return;
+      } catch (Exception e) {
+        LOG.warn("complete file " + src + " failed, retry = " + retry, e);
+      }
+      sleepIgnoreInterrupt(retry);
+    }
+  }
+
+  static void sleepIgnoreInterrupt(int retry) {
+    try {
+      Thread.sleep(ConnectionUtils.getPauseTime(100, retry));
+    } catch (InterruptedException e) {
+    }
+  }
+}

http://git-wip-us.apache.org/repos/asf/hbase/blob/6e9d355b/hbase-server/src/test/java/org/apache/hadoop/hbase/util/FanOutOneBlockAsyncDFSOutputFlushHandler.java
----------------------------------------------------------------------
diff --git a/hbase-server/src/test/java/org/apache/hadoop/hbase/util/FanOutOneBlockAsyncDFSOutputFlushHandler.java b/hbase-server/src/test/java/org/apache/hadoop/hbase/util/FanOutOneBlockAsyncDFSOutputFlushHandler.java
new file mode 100644
index 0000000..cbd0761
--- /dev/null
+++ b/hbase-server/src/test/java/org/apache/hadoop/hbase/util/FanOutOneBlockAsyncDFSOutputFlushHandler.java
@@ -0,0 +1,61 @@
+/**
+ * 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.hbase.util;
+
+import java.nio.channels.CompletionHandler;
+import java.util.concurrent.ExecutionException;
+
+public final class FanOutOneBlockAsyncDFSOutputFlushHandler
+    implements CompletionHandler<Long, Void> {
+
+  private long size;
+
+  private Throwable error;
+
+  private boolean finished;
+
+  @Override
+  public synchronized void completed(Long result, Void attachment) {
+    size = result.longValue();
+    finished = true;
+    notifyAll();
+  }
+
+  @Override
+  public synchronized void failed(Throwable exc, Void attachment) {
+    error = exc;
+    finished = true;
+    notifyAll();
+  }
+
+  public synchronized long get() throws InterruptedException, ExecutionException {
+    while (!finished) {
+      wait();
+    }
+    if (error != null) {
+      throw new ExecutionException(error);
+    }
+    return size;
+  }
+
+  public void reset() {
+    size = 0L;
+    error = null;
+    finished = false;
+  }
+}
\ No newline at end of file

http://git-wip-us.apache.org/repos/asf/hbase/blob/6e9d355b/hbase-server/src/test/java/org/apache/hadoop/hbase/util/TestFanOutOneBlockAsyncDFSOutput.java
----------------------------------------------------------------------
diff --git a/hbase-server/src/test/java/org/apache/hadoop/hbase/util/TestFanOutOneBlockAsyncDFSOutput.java b/hbase-server/src/test/java/org/apache/hadoop/hbase/util/TestFanOutOneBlockAsyncDFSOutput.java
new file mode 100644
index 0000000..0e9f42e
--- /dev/null
+++ b/hbase-server/src/test/java/org/apache/hadoop/hbase/util/TestFanOutOneBlockAsyncDFSOutput.java
@@ -0,0 +1,190 @@
+/**
+ * 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.hbase.util;
+
+import static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_CLIENT_SOCKET_TIMEOUT_KEY;
+import static org.junit.Assert.assertArrayEquals;
+import static org.junit.Assert.assertEquals;
+import static org.junit.Assert.assertTrue;
+import static org.junit.Assert.fail;
+import io.netty.channel.EventLoop;
+import io.netty.channel.EventLoopGroup;
+import io.netty.channel.nio.NioEventLoopGroup;
+
+import java.io.FileNotFoundException;
+import java.io.IOException;
+import java.util.concurrent.ExecutionException;
+import java.util.concurrent.ThreadLocalRandom;
+
+import org.apache.hadoop.fs.FSDataInputStream;
+import org.apache.hadoop.fs.Path;
+import org.apache.hadoop.hbase.HBaseTestingUtility;
+import org.apache.hadoop.hbase.testclassification.MediumTests;
+import org.apache.hadoop.hbase.testclassification.MiscTests;
+import org.apache.hadoop.hdfs.DistributedFileSystem;
+import org.apache.hadoop.ipc.RemoteException;
+import org.apache.log4j.Level;
+import org.apache.log4j.Logger;
+import org.junit.AfterClass;
+import org.junit.BeforeClass;
+import org.junit.Rule;
+import org.junit.Test;
+import org.junit.experimental.categories.Category;
+import org.junit.rules.TestName;
+
+@Category({ MiscTests.class, MediumTests.class })
+public class TestFanOutOneBlockAsyncDFSOutput {
+
+  private static final HBaseTestingUtility TEST_UTIL = new HBaseTestingUtility();
+
+  private static DistributedFileSystem FS;
+
+  private static EventLoopGroup EVENT_LOOP_GROUP;
+
+  private static int READ_TIMEOUT_MS = 2000;
+
+  @Rule
+  public TestName name = new TestName();
+
+  @BeforeClass
+  public static void setUp() throws Exception {
+    Logger.getLogger("org.apache.hadoop.hdfs.StateChange").setLevel(Level.DEBUG);
+    Logger.getLogger("BlockStateChange").setLevel(Level.DEBUG);
+    TEST_UTIL.getConfiguration().setInt(DFS_CLIENT_SOCKET_TIMEOUT_KEY, READ_TIMEOUT_MS);
+    TEST_UTIL.startMiniDFSCluster(3);
+    FS = TEST_UTIL.getDFSCluster().getFileSystem();
+    EVENT_LOOP_GROUP = new NioEventLoopGroup();
+  }
+
+  @AfterClass
+  public static void tearDown() throws IOException, InterruptedException {
+    if (EVENT_LOOP_GROUP != null) {
+      EVENT_LOOP_GROUP.shutdownGracefully().sync();
+    }
+    TEST_UTIL.shutdownMiniDFSCluster();
+  }
+
+  private void writeAndVerify(EventLoop eventLoop, Path f, final FanOutOneBlockAsyncDFSOutput out)
+      throws IOException, InterruptedException, ExecutionException {
+    final byte[] b = new byte[10];
+    ThreadLocalRandom.current().nextBytes(b);
+    final FanOutOneBlockAsyncDFSOutputFlushHandler handler =
+        new FanOutOneBlockAsyncDFSOutputFlushHandler();
+    eventLoop.execute(new Runnable() {
+
+      @Override
+      public void run() {
+        out.write(b, 0, b.length);
+        out.flush(null, handler, false);
+      }
+    });
+    assertEquals(b.length, handler.get());
+    out.close();
+    assertEquals(b.length, FS.getFileStatus(f).getLen());
+    byte[] actual = new byte[b.length];
+    try (FSDataInputStream in = FS.open(f)) {
+      in.readFully(actual);
+    }
+    assertArrayEquals(b, actual);
+  }
+
+  @Test
+  public void test() throws IOException, InterruptedException, ExecutionException {
+    Path f = new Path("/" + name.getMethodName());
+    EventLoop eventLoop = EVENT_LOOP_GROUP.next();
+    final FanOutOneBlockAsyncDFSOutput out =
+        FanOutOneBlockAsyncDFSOutputHelper.createOutput(FS, f, true, false, (short) 3,
+          FS.getDefaultBlockSize(), eventLoop);
+    writeAndVerify(eventLoop, f, out);
+  }
+
+  @Test
+  public void testRecover() throws IOException, InterruptedException, ExecutionException {
+    Path f = new Path("/" + name.getMethodName());
+    EventLoop eventLoop = EVENT_LOOP_GROUP.next();
+    final FanOutOneBlockAsyncDFSOutput out =
+        FanOutOneBlockAsyncDFSOutputHelper.createOutput(FS, f, true, false, (short) 3,
+          FS.getDefaultBlockSize(), eventLoop);
+    final byte[] b = new byte[10];
+    ThreadLocalRandom.current().nextBytes(b);
+    final FanOutOneBlockAsyncDFSOutputFlushHandler handler =
+        new FanOutOneBlockAsyncDFSOutputFlushHandler();
+    eventLoop.execute(new Runnable() {
+
+      @Override
+      public void run() {
+        out.write(b, 0, b.length);
+        out.flush(null, handler, false);
+      }
+    });
+    handler.get();
+    // restart one datanode which causes one connection broken
+    TEST_UTIL.getDFSCluster().restartDataNode(0);
+    handler.reset();
+    eventLoop.execute(new Runnable() {
+
+      @Override
+      public void run() {
+        out.write(b, 0, b.length);
+        out.flush(null, handler, false);
+      }
+    });
+    try {
+      handler.get();
+      fail("flush should fail");
+    } catch (ExecutionException e) {
+      // we restarted one datanode so the flush should fail
+      e.printStackTrace();
+    }
+    out.recoverAndClose(null);
+    assertEquals(b.length, FS.getFileStatus(f).getLen());
+    byte[] actual = new byte[b.length];
+    try (FSDataInputStream in = FS.open(f)) {
+      in.readFully(actual);
+    }
+    assertArrayEquals(b, actual);
+  }
+
+  @Test
+  public void testHeartbeat() throws IOException, InterruptedException, ExecutionException {
+    Path f = new Path("/" + name.getMethodName());
+    EventLoop eventLoop = EVENT_LOOP_GROUP.next();
+    final FanOutOneBlockAsyncDFSOutput out =
+        FanOutOneBlockAsyncDFSOutputHelper.createOutput(FS, f, true, false, (short) 3,
+          FS.getDefaultBlockSize(), eventLoop);
+    Thread.sleep(READ_TIMEOUT_MS * 2);
+    // the connection to datanode should still alive.
+    writeAndVerify(eventLoop, f, out);
+  }
+
+  /**
+   * This is important for fencing when recover from RS crash.
+   */
+  @Test
+  public void testCreateParentFailed() throws IOException {
+    Path f = new Path("/" + name.getMethodName() + "/test");
+    EventLoop eventLoop = EVENT_LOOP_GROUP.next();
+    try {
+      FanOutOneBlockAsyncDFSOutputHelper.createOutput(FS, f, true, false, (short) 3,
+        FS.getDefaultBlockSize(), eventLoop);
+      fail("should fail with parent does not exist");
+    } catch (RemoteException e) {
+      assertTrue(e.unwrapRemoteException() instanceof FileNotFoundException);
+    }
+  }
+}


Mime
View raw message