Skip to content

Latest commit

 

History

History
210 lines (141 loc) · 8.25 KB

TERRAFORM.md

File metadata and controls

210 lines (141 loc) · 8.25 KB
Raw

Getting Started With Terraform For AWS

This document describes how to deploy a AWS Rekognition Service using a combination of AWS resources.

This is a Terraform based installation methodology that reliably automates the complete build, management and destruction processes of all resources. Terraform is an infrastructure-as-code command line tool that will create and configure all of the approximately two dozen software and cloud infrastructure resources that are needed for running the service on AWS infrastructure. These Terraform scripts will install and configure all cloud infrastructure resources and system software on which the service depends. This process will take around 2 minutes to complete and will generate copious amounts of console output.

The service stack consists of the following:

  • a AWS S3 bucket and DynamoDB table for managing Terraform state
  • AWS S3 bucket for storing train and test image sets.
  • DynamoDB Table for persisting Rekognition service results
  • AWS IAM Role for managing service-level role-based security for this service

WARNINGS:

1. This Terraform project will create many AWS resources in other parts of your AWS account including S3, API Gateway, IAM, Rekognition, DynamoDB, CloudWatch and Lambda. You should not directly modify any of these resources, as this could lead to unintended consequences in the safe operation of your Kubernetes cluster up to and including permanent loss of access to the cluster itself.

2. Terraform is a memory intensive application. For best results you should run this on a computer with at least 4Gib of free memory.

I. Installation Prerequisites

Quickstart for Linux & macOS operating systems.

Prerequisite: Obtain an AWS IAM User with administrator privileges, access key and secret key.

Ensure that your environment includes the latest stable releases of the following software packages:

Install required software packages using Homebrew

If necessary, install Homebrew

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
echo 'eval "$(/home/linuxbrew/.linuxbrew/bin/brew shellenv)"' >> /home/ubuntu/.profile
eval "$(/home/linuxbrew/.linuxbrew/bin/brew shellenv)"

Use homebrew to install all required packages.

brew install awscli terraform

Configure the AWS CLI

To configure the AWS CLI run the following command:

aws configure

This will interactively prompt for your AWS IAM user access key, secret key and preferred region.

Setup Terraform

Terraform is a declarative open-source infrastructure-as-code software tool created by HashiCorp. This repo leverages Terraform to create all cloud infrastructure as well as to install and configure all software packages that run inside of Kubernetes. Terraform relies on an S3 bucket for storing its state data, and a DynamoDB table for managing a semaphore lock during operations.

Use these three environment variables for creating the uniquely named resources that the Terraform modules in this repo will be expecting to find at run-time.

IMPORTANT: these three settings should be consistent with the values your set in terraform.tfvars in the next section.

AWS_ACCOUNT=012345678912      # add your 12-digit AWS account number here
$
AWS_REGION=us-east-1          # any valid AWS region code.
AWS_ENVIRONMENT=rekognition   # any valid string. Keep it short -- 3 characters is ideal.

First create an AWS S3 Bucket

AWS_S3_BUCKET="${AWS_ACCOUNT}-tfstate-${AWS_ENVIRONMENT}"

# for buckets created in us-east-1
aws s3api create-bucket --bucket $AWS_S3_BUCKET --region $AWS_REGION

# for all other regions
aws s3api create-bucket --bucket $AWS_S3_BUCKET --region $AWS_REGION --create-bucket-configuration LocationConstraint=$AWS_REGION

Then create a DynamoDB table

AWS_DYNAMODB_TABLE="${AWS_ACCOUNT}-tfstate-lock-${AWS_ENVIRONMENT}"
aws dynamodb create-table --region $AWS_REGION --table-name $AWS_DYNAMODB_TABLE  \
               --attribute-definitions AttributeName=LockID,AttributeType=S  \
               --key-schema AttributeName=LockID,KeyType=HASH --provisioned-throughput  \
               ReadCapacityUnits=1,WriteCapacityUnits=1

II. Build and Deploy

Step 1. Checkout the repository

git clone https://github.com/lpm0073/aws-rekognition.git

Step 2. Configure your Terraform backend

Edit the following snippet so that bucket, region and dynamodb_table are consistent with your values of $AWS_REGION, $AWS_S3_BUCKET, $AWS_DYNAMODB_TABLE

vim terraform/terraform.tf
  backend "s3" {
    bucket         = "012345678912-tfstate-rekognition"
    key            = "rekognition/terraform.tfstate"
    region         = "us-east-1"
    dynamodb_table = "012345678912-tfstate-lock-rekognition"
    profile        = "default"
    encrypt        = false
  }

Step 4. Configure your environment by setting Terraform global variable values

vim terraform/terraform.tfvars

Required inputs are as follows:

account_id           = "012345678912"
aws_region           = "us-east-1"
aws_profile          = "default"

Step 3. Run the following command to initialize and build the solution

The Terraform modules in this repo rely extensively on calls to other third party Terraform modules published and maintained by AWS. These modules will be downloaded by Terraform so that these can be executed locally from your computer. Noteworth examples of such third party modules include:

cd terraform
terraform init

Screen output should resemble the following: Terraform init

terraform plan

Screen output should resemble the following: Terraform plan

To deploy the service run the following

terraform apply

Terraform apply

III. Uninstall

The following completely destroys all AWS resources. Note that this operation might take up to 20 minutes to complete.

cd terraform
terraform init
terraform destroy

Delete Terraform state management resources

AWS_ACCOUNT=012345678912      # add your 12-digit AWS account number here
AWS_REGION=us-east-1
AWS_ENVIRONMENT=rekognition   # any valid string. Keep it short
AWS_S3_BUCKET="${AWS_ACCOUNT}-tfstate-${AWS_ENVIRONMENT}"
AWS_DYNAMODB_TABLE="${AWS_ACCOUNT}-tfstate-lock-${AWS_ENVIRONMENT}"

To delete the DynamoDB table

aws dynamodb delete-table --region $AWS_REGION --table-name $AWS_DYNAMODB_TABLE

To delete the AWS S3 bucket

aws s3 rm s3://$AWS_S3_BUCKET --recursive
aws s3 rb s3://$AWS_S3_BUCKET --force

If You're New To Postman

For your convenience there's a preconfigured 'postman_collection' file added to the root directly of this repo. Regardless of whether you use this template, you'll need to provide the following three pieces of information from the Terraform output:

Postman Configuration

Upload images

Postman Index

Search images

Postman Search