MapReduce Client Report

module.exports = header: 'MapReduce Client Report', handler: ({options}) ->
  properties.read ctx.ssh, "#{options.conf_dir}/mapred-site.xml", (err, config) ->
    return next err if err
    ctx.emit 'report',
      key: 'mapreduce.map.memory.mb'
      value: prink.filesize.from.megabytes config['mapreduce.map.memory.mb']
      raw: config['mapreduce.map.memory.mb']
      default: '1536'
      description: 'Larger resource limit for maps.'
    ctx.emit 'report',
      key: 'mapreduce.map.java.opts'
      value: config['mapreduce.map.java.opts']
      default: '-Xmx1024M'
      description: 'Larger heap-size for child jvms of maps.'
    ctx.emit 'report',
      key: 'mapreduce.reduce.memory.mb'
      value: prink.filesize.from.megabytes config['mapreduce.reduce.memory.mb']
      raw: config['mapreduce.reduce.memory.mb']
      default: '3072'
      description: 'Larger resource limit for reduces.'
    ctx.emit 'report',
      key: 'mapreduce.reduce.java.opts'
      value: config['mapreduce.reduce.java.opts']
      default: '-Xmx2560M'
      description: 'Larger heap-size for child jvms of reduces.'
    ctx.emit 'report',
      key: 'mapreduce.task.io.sort.mb'
      value: prink.filesize.from.megabytes config['mapreduce.task.io.sort.mb']
      raw: config['mapreduce.task.io.sort.mb']
      default: '512'
      description: 'Higher memory-limit while sorting data for efficiency.'
    ctx.emit 'report',
      key: 'mapreduce.task.io.sort.factor'
      value: config['mapreduce.task.io.sort.factor']
      default: '100'
      description: 'More streams merged at once while sorting files.'
    ctx.emit 'report',
      key: 'mapreduce.reduce.shuffle.parallelcopies'
      value: config['mapreduce.reduce.shuffle.parallelcopies']
      default: '50'
      description: 'Higher number of parallel copies run by reduces to fetch outputs from very large number of maps.'
    next null, true

Dependencies

properties = require '../../lib/properties'
prink = require 'prink'