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 2006/05/15 21:21:41 UTC
svn commit: r406722 -
/lucene/hadoop/trunk/src/examples/org/apache/hadoop/examples/Sort.java
Author: cutting
Date: Mon May 15 12:21:40 2006
New Revision: 406722
URL: http://svn.apache.org/viewcvs?rev=406722&view=rev
Log:
Set number of reduce tasks to be the number of reduce nodes. This
results in optimal sort performance. Contributed by Owen.
Modified:
lucene/hadoop/trunk/src/examples/org/apache/hadoop/examples/Sort.java
Modified: lucene/hadoop/trunk/src/examples/org/apache/hadoop/examples/Sort.java
URL: http://svn.apache.org/viewcvs/lucene/hadoop/trunk/src/examples/org/apache/hadoop/examples/Sort.java?rev=406722&r1=406721&r2=406722&view=diff
==============================================================================
--- lucene/hadoop/trunk/src/examples/org/apache/hadoop/examples/Sort.java (original)
+++ lucene/hadoop/trunk/src/examples/org/apache/hadoop/examples/Sort.java Mon May 15 12:21:40 2006
@@ -69,7 +69,7 @@
int num_maps = cluster.getTaskTrackers() *
jobConf.getInt("test.sort.maps_per_host", 10);
int num_reduces = cluster.getTaskTrackers() *
- jobConf.getInt("test.sort.reduces_per_host", 10);
+ jobConf.getInt("test.sort.reduces_per_host", cluster.getMaxTasks());
List otherArgs = new ArrayList();
for(int i=0; i < args.length; ++i) {
try {
@@ -109,7 +109,7 @@
cluster.getTaskTrackers() +
" nodes to sort from " +
jobConf.getInputPaths()[0] + " into " +
- jobConf.getOutputPath() + ".");
+ jobConf.getOutputPath() + " with " + num_reduces + " reduces.");
Date startTime = new Date();
System.out.println("Job started: " + startTime);
JobClient.runJob(jobConf);