Skip to content

Latest commit

 

History

History
83 lines (56 loc) · 4.15 KB

lecture.md

File metadata and controls

83 lines (56 loc) · 4.15 KB
Raw

Cloud Deployment

Material for this Guide

This tutorial assumes a trivial python application, which you may fork from https://github.com/amfleming/python-flask-bootstrap

Amazon ECS Vocabulary

  • Cluster - A set of ECS Instances
  • ECS Instance - An EC2 instance running the ECS agent. A single ECS Instance can run any number of Services.
  • Service - A collection of related containers running together on an ECS Instance. Services are defined by their Task Definition and run on an Instance
  • Task Definition - A defines a service as a set of images and their linkages. Note that a Task Definition in AWS an idential concept (and very similar syntax) to a docker-compose.yml
  • Container - a Docker container
  • ECS Container Agent - An agent running on an ECS Instance whose purpose is to communicate AWS and the Docker daemon running on the instante. This is actually a container itself (!) and open source: [code]

1. Create an empty ECS Cluster

What follows below is covered in much more detail in the The Amazon EC2 Guide

Setup a Key Pair

Without one you will not be able to actually log in to new instances. Go to EC2 -> Key Pairs -> Create. A .pem file will be downloaded to your machine.

To make it easy to ssh into EC2 instances, do the following:

# Keep the key file in your .ssh folder
cp ~/Downloads/MyKeyPair.pem ~/.ssh/

# SSH is very strict!  It will not look at the key file 
# unless you lock it down to only be readable by you
chmod 400 -rw ~/.ssh/MyKeyPair.pem

# Tell ssh to try this key-pair out when connecting to servers
ssh-add ~/.ssh/MyKeyPair.pem

Setup a VPC

  • Setup a new VPC using the VPC Wizard. Accept the defaults
  • Add a Firewall using the Security Groups feature in the VPC. You need SSH and HTTP.
  • Lock down SSH to your IP Address!

Setup a User

Identity & Access Management (IAM) -> New User -> Attach Policy -> PowerUserAccess

2. Build and Push to DockerHub

You application should have a Dockerfile in its root folder which builds a self-contained image of your application. In order for this image to run on AWS, you must push it to docker-hub.

$ docker build .
$ docker tag <image_id> amfleming/python-flask-bootstrap
$ docker push amfleming/python-flask-bootstrap

Troubleshooting

  • You must have a DockerHub account
  • You may need to docker login
  • See the tag, push and pull section of the Docker Tutorial

3. Create a new ECS Task Definition

Run the ECS First-run wizard tool, choosing 'custom'. In addition to creating the new ECS cluster, the wizard will also create a new EC2 instance for you along with an auto-scaling policy. (If you did not use the wizard, you would need to manually launch EC2 Instances configured with the ECS Agent)

Create the Task Definition

Add a single Container Definition with these settings:

  • Container Name can be anything you like
  • Image should be a reference to your docker-hub image. In this case it will be amfleming/python-flask-bootstrap
  • Memory and CPU Units can be anything reasonable
  • Essential should be checked
  • Port Mappings should map port 80 in the container to 80 on the host. If you do not map the port, there will be no errors, but you will not be able to access your python app!

Schedule the Task as a Service

Choose create a service vs Run Tasks Once, to ensure your python web-app is continuously available

Configure Cluster

You'll only need a single host to complete this tutorial, t2.micro is fine for instance type, and choose the Key-Pair that you previously created in step (1). Select the User you created in step (1) as well.

4. Access your Running web App

Once the wizard completes the setup of your cluster, launches an ECS Instance, and deploys your app, you can simply access the url of the EC2 instance (e.g. http://ec2-12-23-234-235.compute-1.amazonaws.com/)