1 name "hadoop_aws_io-bound"
2 description "Hadoop cluster conf. on the AWS EC2 instances for I/O bound processing."
4 #cookbook_versions "hadoop" => "= 0.2.0"
8 'dfs.replication' => '3', # default: 3 (recommended)
11 # e.g. 'core-site.xml' => {'k1' => 'v1', 'k2' => 'v2'},
13 'HADOOP_HEAPSIZE' => '20000', # default: 1000 MB
14 #'HADOOP_NAMENODE_INIT_HEAPSIZE' => '2000' # default: 1000 MB (ver. 2.x)
15 #'HADOOP_NAMENODE_OPTS' => '"${HADOOP_NAMENODE_OPTS} -Xmx20000m"',
16 #'HADOOP_SECONDARYNAMENODE_OPTS' => '"${HADOOP_SECONDARYNAMENODE_OPTS} -Xmx20000m"',
17 #'HADOOP_DATANODE_OPTS' => '"${HADOOP_DATANODE_OPTS} "',
18 #'HADOOP_BALANCER_OPTS' => '"${HADOOP_BALANCER_OPTS} "',
19 #'HADOOP_JOBTRACKER_OPTS' => '"${HADOOP_JOBTRACKER_OPTS} -Xmx20000m"',
21 'yarn-env.sh' => { # ver. 2.x
22 #'YARN_HEAPSIZE' => '20000' # default: 1000 MB
23 #'YARN_RESOURCEMANAGER_OPTS' => '"${YARN_RESOURCEMANAGER_OPTS} -Xmx20000m"',
24 #'YARN_NODEMANAGER_OPTS' => '"${YARN_NODEMANAGER_OPTS} "',
26 'mapred-env.sh' => { # ver. 2.x
27 #'HADOOP_JOB_HISTORYSERVER_HEAPSIZE' => '20000', # default: 1000 MB
28 #'HADOOP_JOB_HISTORYSERVER_OPTS' => '"${HADOOP_JOB_HISTORYSERVER_OPTS} "',
33 'dfs.block.size' => '134217728', # default: 67108864 (bytes)
34 #'dfs.blocksize' => '134217728', # default: 134217728 (bytes) (ver. 2.x)
35 #'dfs.datanode.du.reserved' => '0', # default: 0 (byte)
37 #'dfs.support.append' => 'true',
38 #'dfs.datanode.max.xcievers' => '4096', # default: 4096 (ver. 1.1 or later), 256 (older ver.)
39 #'dfs.datanode.max.transfer.threads' => '4096', # default: 4096 (ver. 2.x)
41 'yarn-site.xml' => { # ver. 2.x
42 #'yarn.nodemanager.resource.cpu-vcores' => '8', # default: 8
43 #'yarn.nodemanager.resource.memory-mb' => '8192', # default: 8192
45 'mapred-site.xml' => {
46 # map tasks:reduce tasks = 4:3 or 2:1 or ...
47 'mapred.tasktracker.map.tasks.maximum' => '8', # default: 2
48 'mapred.tasktracker.reduce.tasks.maximum' => '4', # default: 2
49 #'mapred.child.java.opts' => '-Xmx1000m', # default: -Xmx200m
50 'mapred.reduce.slowstart.completed.maps' => '0.80', # default: 0.05
51 #'mapreduce.job.reduce.slowstart.completedmaps' => '0.80', # default: 0.05 (ver. 2.x)
52 'tasktracker.http.threads' => '80', # default: 40
53 #'mapreduce.tasktracker.http.threads' => '80', # default: 40 (ver. 2.x)
54 #'mapred.reduce.parallel.copies' => SQRT(nodes), # default: 5
55 #'mapreduce.reduce.shuffle.parallelcopies' => SQRT(nodes), # default: 5 (ver. 2.x)
56 'mapred.submit.replication' => '3', # default: 10
57 #'mapreduce.client.submit.file.replication' => '3', # default: 10 (ver. 2.x)
59 'capacity-scheduler.xml' => {}
67 'hadoop' => hadoop_attributes