Do you mean run it on your local machine from Eclipse versus run on a
distributed cluster?
I would guess the Eclipse/Cygwin JVM is 32bit and the cluster JVM is
64bit. References occupy more space on a 64bit JVM. You might want to
check the -XX:+UseCompressedOops VM option.
http://wikis.sun.com/display/HotSpotInternals/CompressedOops
radar.sxl wrote:
When I run a hadoop project with Eclipse-plugin or Cygwin, it's OK. But when
I run in Hadoop cluster, it's error. So strange questions, I do not kown
why!
OK：
09/11/13 15:03:59 INFO mapred.JobClient: Running job: job_200911131503_0001
09/11/13 15:04:01 INFO mapred.JobClient: map 0% reduce 0%
09/11/13 15:04:12 INFO mapred.JobClient: map 50% reduce 0%
09/11/13 15:04:19 INFO mapred.JobClient: map 100% reduce 0%
09/11/13 15:04:24 INFO mapred.JobClient: map 100% reduce 16%
09/11/13 15:04:28 INFO mapred.JobClient: map 100% reduce 100%
09/11/13 15:04:31 INFO mapred.JobClient: Job complete: job_200911131503_0001
Error：
09/11/13 15:17:48 INFO mapred.JobClient: Running job: job_200911131430_0003
09/11/13 15:17:49 INFO mapred.JobClient: map 0% reduce 0%
09/11/13 15:18:20 INFO mapred.JobClient: Task Id :
attempt_200911131430_0003_m_0
00001_0, Status : FAILED
java.lang.OutOfMemoryError: Java heap space
at java.io.BufferedReader.readLine(BufferedReader.java:331)
at java.io.BufferedReader.readLine(BufferedReader.java:362)
at jeasy.analysis.lIlllIIIIIllIlII._$1(Unknown Source:80)
at jeasy.analysis.lIlllIIIIIllIlII._$1(Unknown Source:36)
at jeasy.analysis.lIIllIlIlIIIllll.<clinit>(Unknown Source:12)
at jeasy.analysis.MMAnalyzer.tokenStream(Unknown Source:26)
at jeasy.analysis.lIlllIIIIIllIlII.segment(Unknown Source:234)
at cnWordCount$MapClass.map(cnWordCount.java:66)
at cnWordCount$MapClass.map(cnWordCount.java:1)
at org.apache.hadoop.mapred.MapRunner.run(MapRunner.java:50)
at org.apache.hadoop.mapred.MapTask.run(MapTask.java:342)
at org.apache.hadoop.mapred.Child.main(Child.java:158)
