Bcc'd dev mailing list. Hi Andrew,
The memory allocated will always be an integral multiple of minimal allocation unit which is configured via the property "yarn.scheduler.minimum-allocation-mb ". The default configuration for the value is 1024. If you change the config to 512, the container memory size should be 512 with -Dmapreduce.map.memory.mb=512. Thanks, Chuan -----Original Message----- From: Yuzhang Han [mailto:[email protected]] Sent: Friday, May 31, 2013 12:24 PM To: [email protected]; [email protected] Subject: How to configure container capacity? Hi, Can anyone tell me how to configure the capacity of containers? Which parameters should be set? I use the following command with -Dmapreduce.map.memory.mb=512, but why is the map container size always 1024? $HADOOP_COMMON_HOME/bin/hadoop jar $HADOOP_MAPRED_HOME/share/hadoop/mapreduce/hadoop-mapreduce-examples-2.0.4-alpha.jar randomwriter -Dmapreduce.job.user.name=$USER -Dmapreduce.randomwriter.bytespermap=10000 -Ddfs.block.size=536870912 -Dmapreduce.map.memory.mb=512 output Thank you, Andrew
