ambari

#ambari

####Table of Contents

1. Overview
2. Module Description
3. Setup
   • What this module affects
   • Setup requirements
   • Beginning with Ambari
4. Usage - Configuration options and additional functionality
5. Reference - An under-the-hood peek at what the module is doing and how
6. Limitations - OS compatibility, etc.
7. Development - Guide for contributing to the module

##Overview

This module installs Ambari (the hadoop management and monitoring system) agent and/or server. It is tested with CentOS (6.4).

##Module Description

This modules prepares the hosts for Ambari and installs Agent/Server service. This means that firewall services are stopped, ntp is installed and started, machines, the agent/clients are installed&started and finally the Ambari clients are given information on where to find the server machine.

##Setup

###What this module affects

• IPTables and PackageKIT are stopped by this module as they may interfer with Ambari. For anything but virtual machines you might want to re-activate these.
• If not present, ntp is installed and started.

###Setup Requirements

• This module needs the module stdlib
• Ambari relies on functioning DNS (or: the hosts must know their own hostname and be able to find other hosts by node), which often you won't have (in particular when working with virtual machines). See below for an example on how to handle this.

###Beginning with Ambari

To get started with this module you need to include the following lines in your puppet file to provision the Ambari server:

include ambari
class { 'ambari::server':
  ownhostname => 'one.cluster'
}

For the agent you need to include the following lines:

include ambari
class { 'ambari::agent':

ownhostname => 'two.cluster', serverhostname => 'one.cluster' }

Ownhostname is the name of the host that is provisioned, serverhostname the name of the Ambari server (the server should be running first).

##Usage One possible pitfall is, that you nodes must be able to contact each other using domain names

• something that won't work automatically when you try it out in a cluster of virtual machines. The simplest way to get this to work is to set these names yourself - see below for an example puppet class. This class sets the values necessary for a cluster consisting of just two machines (one.cluster and two.cluster) to find each other. You must include a call to this class like this:

class { 'etchosts': ownhostname => 'two.cluster' }

class etchosts ($ownhostname) { host { 'host_one': name => 'one.cluster', alias => ['one', 'one.cluster'], ip => '192.168.0.101' }

host { 'host_two': name => 'two.cluster', alias => ['two', 'two.cluster'], ip => '192.168.0.102' }

file { 'agent_hostname': path => "/etc/hostname", ensure => present, replace => true, content => "${ownhostname}", # own hostname owner => 1546 }

file { 'agent_sysconfig': path => "/etc/sysconfig/network", ensure => present, replace => true, content => "NETWORKING=yes \nHOSTNAME=${ownhostname}" # own hostname }

}

##Limitations This module is currently tested only on CentOS.

##Development If you have ideas for improvements or want to make changes yourself, please contact us at .@codecentric.de

##Release Notes/Contributors/Etc This Module was created by Malte Nottmeyer and Valentin Zacharias