HOWTO: Deploying Hadoop 2 cluster on OpenStack

This example shows how to deploy a Hadoop 2 cluster on OpenStack with Nuri.

1. clone Nuri

    $ apt-get install git ruby1.9.1 ruby1.9.1-dev libz-dev libaugeas-ruby1.9.1 libxml2-dev libxslt-dev make gcc
    $ gem install sfpagent sfplanner coderay colorize
    $ git clone https://github.com/herry13/nuri
   

2. enter Nuri console mode, and then start local agent (it will become the OpenStack proxy)

    $ cd nuri
    $ ./bin/nuri console
    nuri@user> agent start
   

3. specify the desired state of your system

    nuri@user> edit
   

   type and save the following example specification:

    include "modules/vm/vm.sfp"
    include "modules/openstack/openstack.sfp"
    include "modules/hadoop2/hadoop2.sfp"
    proxy isa Node {
      sfpAddress is "localhost"
      openstack isa OpenStack { 
        auth_uri is "http://16.25.166.21:5000/v2.0/tokens/"
        vm_flavor is "3"
        vm_ssh_key_name is "nurikey"
        vm_image is "8fe14786-5388-4787-8aaf-96feebfa8aae"
      } 
    }
    master isa VM { 
      in_cloud is proxy.openstack 
      hadoop isa Hadoop2Master {
        source is "http://master.nuri.ext9.sup.hpl.hp.com/hadoop"
      }
    } 
    slave1 isa VM { 
      in_cloud is proxy.openstack 
      hadoop isa Hadoop2Slave {
        master is master.hadoop
        source is "http://master.nuri.ext9.sup.hpl.hp.com/hadoop"
      } 
    } 
    slave2 extends slave1
    slave3 extends slave1
    slave4 extends slave1
    slave5 extends slave1
   

4. generate the workflow for testing

    nuri@user> plan
   

5. generate and deploy the BSig model

    nuri@user> bsig -a