HDFS-6123. Do not log stack trace for ReplicaAlreadyExistsException and SocketTimeoutException.

git-svn-id: https://svn.apache.org/repos/asf/hadoop/common/trunk@1579396 13f79535-47bb-0310-9956-ffa450edef68

hadoop-hdfs-project/hadoop-hdfs/CHANGES.txt

@@ -421,6 +421,9 @@ Release 2.4.0 - UNRELEASED
     HDFS-6068. Disallow snapshot names that are also invalid directory names.
     (sathish via szetszwo)
 
+    HDFS-6123. Do not log stack trace for ReplicaAlreadyExistsException and
+    SocketTimeoutException.  (szetszwo)
+
   OPTIMIZATIONS
 
     HDFS-5790. LeaseManager.findPath is very slow when many leases need recovery

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/BlockSender.java

@@ -560,7 +560,11 @@ private int sendPacket(ByteBuffer pkt, int maxChunks, OutputStream out,
          * part of a block and then decides not to read the rest (but leaves
          * the socket open).
          */
-          LOG.info("exception: ", e);
+        if (LOG.isTraceEnabled()) {
+          LOG.trace("Failed to send data:", e);
+        } else {
+          LOG.info("Failed to send data: " + e);
+        }
       } else {
         /* Exception while writing to the client. Connection closure from
          * the other end is mostly the case and we do not care much about

hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/DataXceiver.java

@@ -18,12 +18,12 @@
 package org.apache.hadoop.hdfs.server.datanode;
 
 import static org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.Status.ERROR;
+import static org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.Status.ERROR_ACCESS_TOKEN;
 import static org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.Status.ERROR_INVALID;
 import static org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.Status.ERROR_UNSUPPORTED;
-import static org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.Status.ERROR_ACCESS_TOKEN;
 import static org.apache.hadoop.hdfs.protocol.proto.DataTransferProtos.Status.SUCCESS;
-import static org.apache.hadoop.util.Time.now;
 import static org.apache.hadoop.hdfs.server.datanode.DataNode.DN_CLIENTTRACE_FORMAT;
+import static org.apache.hadoop.util.Time.now;
 
 import java.io.BufferedInputStream;
 import java.io.BufferedOutputStream;
@@ -43,7 +43,6 @@
 import java.util.Arrays;
 
 import org.apache.commons.logging.Log;
-import org.apache.hadoop.fs.InvalidRequestException;
 import org.apache.hadoop.hdfs.ExtendedBlockId;
 import org.apache.hadoop.hdfs.ShortCircuitShm.SlotId;
 import org.apache.hadoop.hdfs.net.Peer;
@@ -51,9 +50,9 @@
 import org.apache.hadoop.hdfs.protocol.ExtendedBlock;
 import org.apache.hadoop.hdfs.protocol.HdfsConstants;
 import org.apache.hadoop.hdfs.protocol.datatransfer.BlockConstructionStage;
+import org.apache.hadoop.hdfs.protocol.datatransfer.DataTransferEncryptor;
 import org.apache.hadoop.hdfs.protocol.datatransfer.DataTransferEncryptor.InvalidMagicNumberException;
 import org.apache.hadoop.hdfs.protocol.datatransfer.DataTransferProtoUtil;
-import org.apache.hadoop.hdfs.protocol.datatransfer.DataTransferEncryptor;
 import org.apache.hadoop.hdfs.protocol.datatransfer.IOStreamPair;
 import org.apache.hadoop.hdfs.protocol.datatransfer.Op;
 import org.apache.hadoop.hdfs.protocol.datatransfer.Receiver;
@@ -231,10 +230,21 @@ public void run() {
       } while ((peer != null) &&
           (!peer.isClosed() && dnConf.socketKeepaliveTimeout > 0));
     } catch (Throwable t) {
-      LOG.error(datanode.getDisplayName() + ":DataXceiver error processing " +
+      String s = datanode.getDisplayName() + ":DataXceiver error processing "
-                ((op == null) ? "unknown" : op.name()) + " operation " +
+          + ((op == null) ? "unknown" : op.name()) + " operation "
-                " src: " + remoteAddress +
+          + " src: " + remoteAddress + " dst: " + localAddress;
-                " dest: " + localAddress, t);
+      if (op == Op.WRITE_BLOCK && t instanceof ReplicaAlreadyExistsException) {
+        // For WRITE_BLOCK, it is okay if the replica already exists since
+        // client and replication may write the same block to the same datanode
+        // at the same time.
+        if (LOG.isTraceEnabled()) {
+          LOG.trace(s, t);
+        } else {
+          LOG.info(s + "; " + t);
+        }
+      } else {
+        LOG.error(s, t);
+      }
     } finally {
       if (LOG.isDebugEnabled()) {
         LOG.debug(datanode.getDisplayName() + ":Number of active connections is: "