Basic Monitoring for Hadoop Data Nodes

Here’s a basic monitoring script to monitor the HDFS cluster disk space, Temp Dir space and number of data nodes up. This was plenty useful before we switched to Cloudera Manager.

#!/usr/bin/env ruby

# Checks Hadoop and alerts if there is a change in data nodes
require 'yaml'

DFS_USED_PERCENT_THRESHOLD = 75
error_flag = 0
report = `hdfs dfsadmin -report`

# Datanodes available: 17 (28 total, 11 dead)
datanodes_report = report[/Data.*/]
persistence_report = {}
old_persistence_report = {}
dfs_used_percent = report[/DFS Used%.*/].split(' ')[-1].to_i

datanodes_total = datanodes_report[/\d+\stotal/].split(' ')[0]
datanodes_dead = datanodes_report[/\d+\sdead/].split(' ')[0]
datanodes_available = datanodes_report[/available:\s\d+/].split(' ')[-1]

persistence_report['datanodes_total'] = datanodes_total
persistence_report['datanodes_dead'] = datanodes_dead
persistence_report['datanodes_available'] = datanodes_available

# Load report from previous run
old_persistence_report = YAML::load_file("/etc/hadoop_report/hadoop_report.yml")


# Persist hash in a file
File.open("/etc/hadoop_report/hadoop_report.yml", "w") do |file|
  file.write persistence_report.to_yaml
end

# Node monitoring
puts "New Report"
puts persistence_report.each {|k,v| puts "#{k} = #{v}"}
p;p;
puts "Old Report"
puts old_persistence_report.each {|k,v| puts "#{k} = #{v}"}

# Storage monitoring
puts "Threshold = #{DFS_USED_PERCENT_THRESHOLD}%"
puts "DFS Used% = #{dfs_used_percent}%"
tmp_dir_report = `hdfs dfs -du -s /tmp`
puts "Temp Directory Space = #{tmp_dir_report.split(' ')[0].to_f/(1024)**4} TB"

if persistence_report != old_persistence_report
  puts "There was a change in the Number of Data Nodes available."

  error_flag = 1
end

if dfs_used_percent > DFS_USED_PERCENT_THRESHOLD
  puts "DFS Storage Space Used Threshold Exceeded"

  error_flag = 1
end

exit 1 if error_flag == 1

puts "Everything is OK"
 
6
Kudos
 
6
Kudos

Now read this

Hot Swapping of Tables in Hive

Here are the steps to hot swap a table in Hive when you manually load data in them. Hot Swapping the table allows you to refresh an entire table without any downtime thus making your table always available for querying. With this... Continue →