terraform/examples/azure-spark-and-cassandra-o.../README.md

3.9 KiB

Spark & Cassandra on CentOS 7.x

This Terraform template was based on this Azure Quickstart Template. Changes to the ARM template that may have occurred since the creation of this example may not be reflected here.

This project configures a Spark cluster (1 master and n-slave nodes) and a single node Cassandra on Azure using CentOS 7.x. The base image starts with CentOS 7.3, and it is updated to the latest version as part of the provisioning steps.

Please note that [Azure Resource Manager][3] is used to provision the environment.

Software

Category Software Version Notes
Operating System CentOS 7.x Based on CentOS 7.1 but it will be auto upgraded to the lastest point release
Java OpenJDK 1.8.0 Installed on all servers
Spark Spark 1.6.0 with Hadoop 2.6 The installation contains libraries needed for Hadoop 2.6
Cassandra Cassandra 3.2 Installed through DataStax's YUM repository

Defaults

Component Setting Default Notes
Spark - Master VM Size Standard D1 V2
Spark - Master Storage Standard LRS
Spark - Master Internal IP 10.0.0.5
Spark - Master Service User Account spark Password-less access
Spark - Slave VM Size Standard D3 V2
Spark - Slave Storage Standard LRS
Spark - Slave Internal IP Range 10.0.1.5 - 10.0.1.255
Spark - Slave # of Nodes 2 Maximum of 200
Spark - Slave Availability 2 fault domains, 5 update domains
Spark - Slave Service User Account spark Password-less access
Cassandra VM Size Standard D3 V2
Cassandra Storage Standard LRS
Cassandra Internal IP 10.2.0.5
Cassandra Service User Account cassandra Password-less access

Prerequisites

  1. Ensure you have an Azure subscription.
  2. Ensure you have enough available vCPU cores on your subscription. Otherwise, you will receive an error during the process. The number of cores can be increased through a support ticket in Azure Portal.

main.tf

The main.tf file contains the actual resources that will be deployed. It also contains the Azure Resource Group definition and any defined variables.

outputs.tf

This data is outputted when terraform apply is called, and can be queried using the terraform output command.

provider.tf

Azure requires that an application is added to Azure Active Directory to generate the client_id, client_secret, and tenant_id needed by Terraform (subscription_id can be recovered from your Azure account details). Please go here for full instructions on how to create this to populate your provider.tf file.

terraform.tfvars

If a terraform.tfvars file is present in the current directory, Terraform automatically loads it to populate variables. We don't recommend saving usernames and password to version control, but you can create a local secret variables file and use -var-file to load it.

If you are committing this template to source control, please insure that you add this file to your .gitignore file.

variables.tf

The variables.tf file contains all of the input parameters that the user can specify when deploying this Terraform template.

Post-Deployment

  1. All servers will have a public IP and SSH port enabled by default. These can be disabled or modified in the template or by using Azure Portal.
  2. All servers are configured with the same username and password. You may SSH into each server and ensure connectivity.
  3. Spark WebUI is running on port 8080. Access it using MASTER_WEB_UI_PUBLIC_IP:8080 on your browser. Public IP is available in the outputs as well as through Azure Portal.
  4. Delete the Resource Group that was created to stage the provisioning scripts.