1 name "hadoop_aws_mem-bound"
2 description "Hadoop cluster conf. on the AWS EC2 instances for Memory bound processing."
4 #cookbook_versions "hadoop" => "= 0.2.0"
8 'dfs.replication' => '2',
9 #'dfs.replication' => '3', # default, recommended.
12 # e.g. 'core-site.xml' => {'k1' => 'v1', 'k2' => 'v2'},
14 #'HADOOP_HEAPSIZE' => '2000', # default: 1000 MB
15 #'HADOOP_NAMENODE_INIT_HEAPSIZE' => '2000' # default: 1000 MB (ver. 2.x)
16 #'HADOOP_NAMENODE_OPTS' => '"${HADOOP_NAMENODE_OPTS} -Xmx8000m"',
17 #'HADOOP_SECONDARYNAMENODE_OPTS' => '"${HADOOP_SECONDARYNAMENODE_OPTS} -Xmx8000m"',
18 #'HADOOP_DATANODE_OPTS' => '"${HADOOP_DATANODE_OPTS} "',
19 #'HADOOP_BALANCER_OPTS' => '"${HADOOP_BALANCER_OPTS} "',
20 #'HADOOP_JOBTRACKER_OPTS' => '"${HADOOP_JOBTRACKER_OPTS} -Xmx8000m"',
22 'yarn-env.sh' => { # ver. 2.x
23 #'YARN_HEAPSIZE' => '2000' # default: 1000 MB
24 #'YARN_RESOURCEMANAGER_OPTS' => '"${YARN_RESOURCEMANAGER_OPTS} -Xmx8000m"',
25 #'YARN_NODEMANAGER_OPTS' => '"${YARN_NODEMANAGER_OPTS} "',
27 'mapred-env.sh' => { # ver. 2.x
28 #'HADOOP_JOB_HISTORYSERVER_HEAPSIZE' => '2000', # default: 1000 MB
29 #'HADOOP_JOB_HISTORYSERVER_OPTS' => '"${HADOOP_JOB_HISTORYSERVER_OPTS} "',
34 'dfs.block.size' => '134217728', # default: 67108864 (bytes)
35 #'dfs.blocksize' => '134217728', # default: 134217728 (bytes) (ver. 2.x)
36 #'dfs.datanode.du.reserved' => '0', # default: 0 (byte)
38 #'dfs.support.append' => 'true',
39 #'dfs.datanode.max.xcievers' => '4096', # default: 4096 (ver. 1.1 or later), 256 (older ver.)
40 #'dfs.datanode.max.transfer.threads' => '4096', # default: 4096 (ver. 2.x)
42 'yarn-site.xml' => { # ver. 2.x
43 #'yarn.nodemanager.resource.cpu-vcores' => '8', # default: 8
44 #'yarn.nodemanager.resource.memory-mb' => '8192', # default: 8192
46 'mapred-site.xml' => {
47 # map tasks:reduce tasks = 4:3 or 2:1 or ...
48 'mapred.tasktracker.map.tasks.maximum' => '2', # default: 2
49 'mapred.tasktracker.reduce.tasks.maximum' => '1', # default: 2
50 #'mapred.child.java.opts' => '-Xmx1000m', # default: -Xmx200m
51 'mapred.reduce.slowstart.completed.maps' => '0.80', # default: 0.05
52 #'mapreduce.job.reduce.slowstart.completedmaps' => '0.80', # default: 0.05 (ver. 2.x)
53 #'tasktracker.http.threads' => '80', # default: 40
54 #'mapreduce.tasktracker.http.threads' => '80', # default: 40 (ver. 2.x)
55 #'mapred.reduce.parallel.copies' => SQRT(nodes), # default: 5
56 #'mapreduce.reduce.shuffle.parallelcopies' => SQRT(nodes), # default: 5 (ver. 2.x)
57 'mapred.submit.replication' => '2', # default: 10
58 #'mapreduce.client.submit.file.replication' => '3', # default: 10 (ver. 2.x)
60 'capacity-scheduler.xml' => {}
68 'hadoop' => hadoop_attributes