You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-commits@hadoop.apache.org by cu...@apache.org on 2007/07/19 23:38:41 UTC

svn commit: r557782 - in /lucene/hadoop/branches/branch-0.14: CHANGES.txt src/java/org/apache/hadoop/dfs/DataNode.java

Author: cutting
Date: Thu Jul 19 14:38:40 2007
New Revision: 557782

URL: http://svn.apache.org/viewvc?view=rev&rev=557782
Log:
Merge -r 557779:557780 from trunk to 0.14 branch.  Fixes: HADOOP-1625.

Modified:
    lucene/hadoop/branches/branch-0.14/CHANGES.txt
    lucene/hadoop/branches/branch-0.14/src/java/org/apache/hadoop/dfs/DataNode.java

Modified: lucene/hadoop/branches/branch-0.14/CHANGES.txt
URL: http://svn.apache.org/viewvc/lucene/hadoop/branches/branch-0.14/CHANGES.txt?view=diff&rev=557782&r1=557781&r2=557782
==============================================================================
--- lucene/hadoop/branches/branch-0.14/CHANGES.txt (original)
+++ lucene/hadoop/branches/branch-0.14/CHANGES.txt Thu Jul 19 14:38:40 2007
@@ -376,6 +376,9 @@
 117. HADOOP-1620.  Reduce the number of abstract FileSystem methods,
      simplifying implementations.  (cutting)
 
+118. HADOOP-1625.  Fix a "could not move files" exception in datanode.
+     (Raghu Angadi via cutting)
+
 
 Release 0.13.0 - 2007-06-08
 

Modified: lucene/hadoop/branches/branch-0.14/src/java/org/apache/hadoop/dfs/DataNode.java
URL: http://svn.apache.org/viewvc/lucene/hadoop/branches/branch-0.14/src/java/org/apache/hadoop/dfs/DataNode.java?view=diff&rev=557782&r1=557781&r2=557782
==============================================================================
--- lucene/hadoop/branches/branch-0.14/src/java/org/apache/hadoop/dfs/DataNode.java (original)
+++ lucene/hadoop/branches/branch-0.14/src/java/org/apache/hadoop/dfs/DataNode.java Thu Jul 19 14:38:40 2007
@@ -946,6 +946,14 @@
               // Write checksum
               checksumOut.write( buf, len, checksumSize );
               myMetrics.wroteBytes( len );
+            } else {
+              /* Should we sync() files here? It can add many millisecs of
+               * latency. We did not sync before HADOOP-1134 either.
+               */ 
+              out.close();
+              out = null;
+              checksumOut.close();
+              checksumOut = null;
             }
             
           } catch (IOException iex) {