Community Articles

Find and share helpful community-sourced technical articles.
avatar
Super Guru

Objective

Deploy a 4-node HDP 2.4.2 cluster with Apache Ambari 2.2.2, Vagrant and VirtualBox on OS X host. This is helpful for development and proof of concepts.

Scope

This approach has been tested on OS X host, but it should work on all supported Vagrant and VirtualBox environments.

Pre-requisites

git clone https://github.com/cstanca1/hdp2_4_2-vagrant.git

Create and Start VMs

Change directory to /hdp_2.4.2-vagrant, the folder that includes Vagrantfile and create a /data folder:

mkdir data

This /data folder will be needed for guest VMs to share with the host.

Vagrant (via Vagrantfile) is configured to use Centos 6.7 as the base box and includes the pre-requisites for installing HDP. 4 VMs will be created: 1 Ambari Server (ambari1), 1 Hadoop master (master1) and 2 slaves (slave1, slave2).

vagrant up ambari1

Install and Setup Ambari Server

Set a Local Reference to a Remote Ambari Repo

vagrant ssh ambari1

sudo su -

cd /etc/yum.repos.d

wget http://public-repo-1.hortonworks.com/ambari/centos6/2.x/updates/2.2.2.0/ambari.repo

Setup SSH Access and Starting the Other 3 VMs

Add at the same path with files downloaded from the repoosity, your id_rsa and id_rsa.pub keys (seehttps://wiki.centos.org/HowTos/Network/SecuringSSH section 7 for instructions on CentOS). You could perform these steps on ambari1 VM and copy these two files to your /vagrant_data folder which shares data between guest and host. Only after you copy those two files, start the other three VMs:

vagrant up master1

vagrant up slave1

vagrant up slave2

Install Ambari Server

yum install ambari-server

Setup Ambari Server

Run the setup command to configure your Ambari Server, Database, JDK, LDAP, and other options:

ambari-server setup

Start Ambari Server

ambari-server start

Deploy Cluster using Ambari Web UI

Open up a web browser and go to:

http://ambari1:8080

Log in with username admin and password admin and follow on-screen instructions, using hosts created and selecting services of interest. For more details, see "Automated Install" at: https://docs.hortonworks.com/HDPDocuments/Ambari/Ambari-2.2.2.0/index.html

8,448 Views
Comments

For full guide on the ambari quickstart on vagrant follow the apache doc: https://cwiki.apache.org/confluence/display/AMBARI/Quick+Start+Guide

avatar
Expert Contributor

This is great. It will save me lots of time. I am trying this on an Ubuntu host, not a Mac, and everything is fine until I get down to trying to access ambari1 through the web. I don't have ambari1 in DNS anywhere.

I can "vagrant ssh ambari1" and find the IP address, but presumably that wont let me install without FQDNs.

Any ideas? Thanks again.

avatar
Expert Contributor

To partly answer my own question... Hadoop doesn't need DNS if all the machines are already in the hosts file. So at some point in the above the /etc/hosts file was populated with

192.168.0.11 ambari1.mycluster ambari1

192.168.0.12 master1.mycluster master1

192.168.0.21 slave1.mycluster slave1

192.168.0.22 slave2.mycluster slave2

avatar
Expert Contributor

You use the phrase "Create a Local Ambari Repository". that is not what you are doing. You are merely pointing you machines to look at an external repository. It would be great if you did create a local ambari repository - but that would require a lot more explanation. I had a quick look at Nexus for this, but found it was not trivial to set up for my purposes.

@Alex McLintock

You are correct. This is a local reference to a remote repo. Here is a reference for an actual local repo with no Internet access, but it gets complicated with Vagrant: https://docs.hortonworks.com/HDPDocuments/Ambari-2.2.2.0/bk_Installing_HDP_AMB/content/_setting_up_a...