svn commit: r1461615 - in /hadoop/common/branches/branch-1: CHANGES.txt src/hdfs/org/apache/hadoop/hdfs/server/datanode/DataNode.java

[ki...@apache.org]

Author: kihwal
Date: Wed Mar 27 14:45:14 2013
New Revision: 1461615

URL: http://svn.apache.org/r1461615
Log:
HDFS-4581. DataNode.checkDiskError should not be called on network errors. Contributed by Rohit Kochar.

Modified:
    hadoop/common/branches/branch-1/CHANGES.txt
    hadoop/common/branches/branch-1/src/hdfs/org/apache/hadoop/hdfs/server/datanode/DataNode.java

Modified: hadoop/common/branches/branch-1/CHANGES.txt
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-1/CHANGES.txt?rev=1461615&r1=1461614&r2=1461615&view=diff
==============================================================================
--- hadoop/common/branches/branch-1/CHANGES.txt (original)
+++ hadoop/common/branches/branch-1/CHANGES.txt Wed Mar 27 14:45:14 2013
@@ -28,6 +28,9 @@ Release 1.3.0 - unreleased
     HDFS-4622. Remove redundant synchronized from 
     FSNamesystem#rollEditLog in branch-1. (Jing Zhao via suresh)
 
+    HDFS-4581. DataNode.checkDiskError should not be called on network errors.
+    (Rohit Kochar via kihwal)
+
 Release 1.2.0 - unreleased
 
   INCOMPATIBLE CHANGES

Modified: hadoop/common/branches/branch-1/src/hdfs/org/apache/hadoop/hdfs/server/datanode/DataNode.java
URL: http://svn.apache.org/viewvc/hadoop/common/branches/branch-1/src/hdfs/org/apache/hadoop/hdfs/server/datanode/DataNode.java?rev=1461615&r1=1461614&r2=1461615&view=diff
==============================================================================
--- hadoop/common/branches/branch-1/src/hdfs/org/apache/hadoop/hdfs/server/datanode/DataNode.java (original)
+++ hadoop/common/branches/branch-1/src/hdfs/org/apache/hadoop/hdfs/server/datanode/DataNode.java Wed Mar 27 14:45:14 2013
@@ -27,8 +27,10 @@ import java.io.OutputStream;
 import java.net.InetSocketAddress;
 import java.net.ServerSocket;
 import java.net.Socket;
+import java.net.SocketException;
 import java.net.SocketTimeoutException;
 import java.net.UnknownHostException;
+import java.nio.channels.ClosedByInterruptException;
 import java.nio.channels.ServerSocketChannel;
 import java.nio.channels.SocketChannel;
 import java.security.NoSuchAlgorithmException;
@@ -923,10 +925,16 @@ public class DataNode extends Configured
   /** Check if there is no space in disk 
    *  @param e that caused this checkDiskError call
    **/
-  protected void checkDiskError(Exception e ) throws IOException {
-    
-    LOG.warn("checkDiskError: exception: ", e);
-    
+  protected void checkDiskError(Exception e ) throws IOException {    
+    LOG.warn("checkDiskError: exception: ", e);  
+    if (e instanceof SocketException || e instanceof SocketTimeoutException
+    	  || e instanceof ClosedByInterruptException 
+    	  || e.getMessage().startsWith("Broken pipe")) {
+      LOG.info("Not checking disk as checkDiskError was called on a network" +
+        " related exception");	
+      return;
+    }
+
     if (e.getMessage() != null &&
         e.getMessage().startsWith("No space left on device")) {
       throw new DiskOutOfSpaceException("No space left on device");
@@ -1543,8 +1551,11 @@ public class DataNode extends Configured
         LOG.warn(dnRegistration + ":Failed to transfer " + b + " to " + targets[0].getName()
             + " got " + StringUtils.stringifyException(ie));
         // check if there are any disk problem
-        datanode.checkDiskError();
-        
+        try{
+          checkDiskError(ie);
+        } catch(IOException e) {
+          LOG.warn("DataNode.checkDiskError failed in run() with: ", e);
+        }        
       } finally {
         xmitsInProgress.getAndDecrement();
         IOUtils.closeStream(blockSender);