Part3 - smk's notes

Notes : Docker¶

Section 8 : Building a Multi-Container Application¶

Single Container Deployment Issues¶

The app was simple - no outside dependencies
Our image was built multiple times
How do we connect to a database from a container

We gonna build an app fibbonacci number calculator but a very complex one :)

Architecture¶

If user visits our webpage, and response will be handled by nginx, which will forward request to either React or Express Server which in turn will communicate with Redis and PostgresSQL Server. Redis Server will have a Worker.

Redis is for a in memory, temporary storage while PostgreSQL will be for permanent storage.

Download Entire Project Here Link . Then checkout branch checkpoint-1

Worker Process : Watches redis for new indice. Pulls each new indice, Calculate the appropriate fibbonaci value for it.

Section 9 : “Dockerizing ” Multiple Services¶

Dockerizing a React App¶

Create a Dockerfile.dev in all three services. Create the following Dockerfile.dev in client folder.

FROM node:16-alpine
WORKDIR '/app'
COPY package.json .
RUN npm install
COPY . .
CMD [ "npm","run", "start" ]

Build : docker build -d Dockerfile.dev .

Similarly create the same file in server , worker and change the last instruction to CMD [ "npm", "run", "dev"] and after that build the docker image in each of them.

Designing docker-compose file

postgress
- Image
redist
- Image
server
- specify build
- specify volumes
- Specify env variables

Make a docker-compose.yml to main root working directory.

Note : if we put Environment Variable :

variableName=value : Sets a variable in the container at run time
variableName : Value is taken from your machine at the run time.

version: '3'
services:
  postgres:
    image: postgres:latest
    environment:
      - POSTGRES_PASSWORD=mysecretpassword
  redis:
    image: redis:latest
  api:
    build:
      dockerfile: Dockerfile.dev
      context: ./server/
    volumes:
      - /app/node_modules
      - ./server:/app
    environment:
      - PGUSER=postgres
      - PGHOST=postgres
      - PGDATABASE=postgres
      - PGPASSWORD=postgres_password
      - PGPORT=5432
      - REDIS_HOST=redis
      - REDIS_PORT=6379
      - REDIS_PASSWORD=redis_password

  client:
    build:
      dockerfile: Dockerfile.dev
      context: ./client/
    volumes:
      - /app/node_modules
      - ./client:/app
    ports:
      - "8080:8080"
    environment:
      - REDIS_HOST=redis
      - REDIS_PORT=6379
      - REDIS_PASSWORD=redis_password

  worker:
    build:
      dockerfile: Dockerfile.dev
      context: ./worker/
    volumes:
      - /app/node_modules
      - ./worker:/app

Routing with Nginx¶

So our nginx server will look at the request and properly forward it to either React server or Express Server. A natural question arises that why did not we hosted React and Express Server separately on different ports and then just made reference using port. Usually its not recommended to use static PORTS in address because ports change all the time at different environments and for changing such a small thing we might need to rebuild our apps again.

Nginx reads default.conf for implementing routing.

General Design :

Tell Nginx that there is a upstream server at client:3000
Tell Nginx that there is a upstream server at server:5000
Listen on port 80
If anyone come to / send them to client upstream
If anyone comes /api send them to client upstream

upstream client {
    server client:3000;
}

upstream api {
    server api:5000;
}

server {
    listen 80;
    location / {
        proxy_pass http://client;
    }
    location /api {
            # Rewrite rule changes /api to / for api server
        rewrite ^/api/(.*)$ /$1 break;
        proxy_pass http://api;
    }
}

NOTE : edit docker-compose.yml to change server with api because we changed name to fix confusion while writing nginx rules.

Building a Custom nginx Image¶

Add this Dockerfile.dev to nginx folder.

FROM nginx
COPY ./default.conf /etc/nginx/conf.d/default.conf

And add this to docker-compose.yml

nginx:
  restart: always
  build:
    dockerfile: Dockerfile.dev
    context: ./nginx
  ports:
    - '3000:80'
  depends_on:
    - api
    - client

Starting Docker Compose¶

Be Carefull with dependencies in docker-compose. for example server/api and client always need redis running.

docker-compose up --build

Websocket fix¶

To fix websocket issue with react do this to docker-compose.yml

  client:
    environment:
      - WDS_SOCKET_PORT=0

and to ngnix.conf

location /ws {
      proxy_pass http://client;
      proxy_http_version 1.1;
      proxy_set_header Upgrade $http_upgrade;
      proxy_set_header Connection "Upgrade";
}

Link to complete setup Link : Use the branch checkpoint-2

Section 10 : A Continuos Integration Workflow for Multiple Containers¶

Production Multi-Container Deployments¶

MultiContainer Setup

Push code to github
Travis automatically pulls repo
Travis builds a test image, tests code
Travis builds prod images
Travis pushes built prod images to Docker Hub
Travis pushes built prod images to Docker Hub
Travis pushes project to AWS EB
EB pulls images from DockerHub, deploys

So Advantage of this setup is we are not relying on EB to build our image instead we put our image on dockerhub which in turn can be used by AWS EB or Google Cloud for updating our website.

Create Dockerfile for prod in all four services and replace CMD ["npm", "run", "dev"] in Dockerfile.dev with CMD ["npm","run","start"] in Dockerfile.

Multiple Nginx Instances¶

In case of our Single Container App we had nginx running with prod files which were built by npm build. But now we can replace Client/React Frontend with similar nginx configuration. So now basically we have 2 ngnix servers.

One nginx server focuses on only routing and second one will only serve production file.

Can we replace both ngnix server with only ?? YES :) But we might don’t want to do that for now just because its very common in real life deployments that we might have to manage multiple ngnix servers.

Altering Ngnix’s Listen Port¶

Create nginx folder inside the clien folder and create default.conf.

Create default.conf with contents

server {
  listen 3000;

  location / {
    root /usr/share/nginx/html;
    index index.html index.htm;
    try_files $uri $uri/ /index.html;
  }
}

For Dockerfile

FROM node:16-alpine
WORKDIR '/app'
COPY package.json .
RUN npm install
COPY . .
RUN npm run build

FROM nginx
EXPOSE 80
COPY ./nginx/default.conf /etc/nginx/conf.d/default.conf
COPY --from=builder /app/build /usr/share/ngnix/html

NOTE : here we wrote ngnix.conf only because our frontend is supposed to be hosted on port : 3000.

NOTE : Clean up App.test.js file’s tests they might crash our frontend (removed the test).

Checkout Files at checkpoint-3 link

Setup Github Repo and Similar Travis CI Repo¶

Travis Configuration Setup¶

language: generic
sudo: required
services:
  - docker

before_install:
  - docker build -t smk781/react-test -f ./client/Dockerfile.dev ./client/

script:
  - docker run -e CI=true USERNAME/react-test npm test

after_success:
  - docker build -t smk781/multi-client ./client
  - docker build -t smk781/multi-server ./server
  - docker build -t smk781/multi-nginx ./nginx
  - docker build -t smk781/worker ./worker

Pushing Images to Docker Hub¶

Add these lines below the after_sucess to push existing images to Docker Hub.

after_success:
  # Log in to the Docker CLI
  echo "$DOCKER_PASSWORD" | docker login -u "$DOCKER_USERNAME" --password-stdin
  # Take those images and push them to Docker Hub
  docker push smk781/multi-client
  docker push smk781/multi-server
  docker push smk781/multi-nginx
  docker push smk781/worker
  # Log out of the Docker CLI
  docker logout

Section 11: Multi-Container Deployments to AWS¶

NOTE : Docker running on 64 bit Amazon Linux will work only till July, 2022. So for Amazon Linux 2 Steps are dramatically different.

We will focus only on Amazon Linux 2.

Also notice we won’t be having our normal redis and postgresql containers because we are going to use services offered by Amazon, AWS Elastic Cache and AWS Relational Database Service.

A Big reason use those is they are better configured, more secure, automatically scalable, (RDS) automated backups and rollbacks, logging + maintenance and Easy to migrated off the EB.

EBS Application Creation (If using Multi-Container Docker Platform)¶

Go to AWS Management Console and use Find Services to search for Elastic Beanstalk
Click “Create Application”
Set Application Name to 'multi-docker'
Scroll down to Platform and select Docker
In Platform Branch, select Multi-Container Docker running on 64bit Amazon Linux
Click Create Application
You may need to refresh, but eventually, you should see a green checkmark underneath Health.

EBS Application Creation (If using Amazon Linux 2 Platform Platform)¶

Make sure you have followed the guidance in this note.
Go to AWS Management Console and use Find Services to search for Elastic Beanstalk
Click “Create Application”
Set Application Name to 'multi-docker'
Scroll down to Platform and select Docker
The Platform Branch should be automatically set to Docker Running on 64bit Amazon Linux 2.
Click Create Application
You may need to refresh, but eventually, you should see a green checkmark underneath Health.

RDS Database Creation¶

Go to AWS Management Console and use Find Services to search for RDS
Click Create database button
Select PostgreSQL
Change Version to the newest available v12 version (The free tier is currently not available for Postgres v13)
In Templates, check the Free tier box.
Scroll down to Settings.
Set DB Instance identifier to multi-docker-postgres
Set Master Username to postgres
Set Master Password to postgrespassword and confirm.
Scroll down to Connectivity. Make sure VPC is set to Default VPC
Scroll down to Additional Configuration and click to unhide.
Set Initial database name to fibvalues
Scroll down and click Create Database button

ElastiCache Redis Creation¶

Go to AWS Management Console and use Find Services to search for ElastiCache
Click Redis in sidebar
Click the Create button
Make sure Cluster Mode Enabled is NOT ticked
In Redis Settings form, set Name to multi-docker-redis
Change Node type to 'cache.t2.micro'
Change Replicas per Shard to 0
Scroll down and click Create button

*Creating a Custom Security Group*

Go to AWS Management Console and use Find Services to search for VPC
Find the Security section in the left sidebar and click Security Groups
Click Create Security Group button
Set Security group name to multi-docker
Set Description to multi-docker
Make sure VPC is set to default VPC
Scroll down and click the Create Security Group button.
After the security group has been created, find the Edit inbound rules button.
Click Add Rule
Set Port Range to 5432-6379
Click in the box next to Source and start typing 'sg' into the box. Select the Security Group you just created.
Click the Save rules button

Applying Security Groups to ElastiCache¶

Go to AWS Management Console and use Find Services to search for ElastiCache
Click Redis in Sidebar
Check the box next to Redis cluster
Click Actions and click Modify
Click the pencil icon to edit the VPC Security group. Tick the box next to the new multi-docker group and click Save
Click Modify

*Applying Security Groups to RDS*

Go to AWS Management Console and use Find Services to search for RDS
Click Databases in Sidebar and check the box next to your instance
Click Modify button
Scroll down to Connectivity and add the new multi-docker security group
Scroll down and click the Continue button
Click Modify DB instance button

*Applying Security Groups to Elastic Beanstalk*

Go to AWS Management Console and use Find Services to search for Elastic Beanstalk
Click Environments in the left sidebar.
Click MultiDocker-env
Click Configuration
In the Instances row, click the Edit button.
Scroll down to EC2 Security Groups and tick box next to multi-docker
Click Apply and Click Confirm
After all the instances restart and go from No Data to Severe, you should see a green checkmark under Health.

Add AWS configuration details to .travis.yml file's deploy script¶

Set the region. The region code can be found by clicking the region in the toolbar next to your username. eg: 'us-east-1'
app should be set to the EBS Application Name eg: 'multi-docker'
env should be set to your EBS Environment name. eg: 'MultiDocker-env'
Set the bucket_name. This can be found by searching for the S3 Storage service. Click the link for the elasticbeanstalk bucket that matches your region code and copy the name.
eg: 'elasticbeanstalk-us-east-1-923445599289'
Set the bucket_path to 'docker-multi'
Set access_key_id to $AWS_ACCESS_KEY
Set secret_access_key to $AWS_SECRET_KEY

Setting Environment Variables¶

Go to AWS Management Console and use Find Services to search for Elastic Beanstalk
Click Environments in the left sidebar.
Click MultiDocker-env
Click Configuration
In the Software row, click the Edit button
Scroll down to Environment properties
In another tab Open up ElastiCache, click Redis and check the box next to your cluster. Find the Primary Endpoint and copy that value but omit the :6379
Set REDIS_HOST key to the primary endpoint listed above, remember to omit :6379
Set REDIS_PORT to 6379
Set PGUSER to postgres
Set PGPASSWORD to postgrespassword
In another tab, open up the RDS dashboard, click databases in the sidebar, click your instance and scroll to Connectivity and Security. Copy the endpoint.
Set the PGHOST key to the endpoint value listed above.
Set PGDATABASE to fibvalues
Set PGPORT to 5432
Click Apply button
After all instances restart and go from No Data, to Severe, you should see a green checkmark under Health.

IAM Keys for Deployment¶

You can use the same IAM User's access and secret keys from the single container app we created earlier, or, you can create a new IAM user for this application:

Search for the "IAM Security, Identity & Compliance Service"
Click "Create Individual IAM Users" and click "Manage Users"
Click "Add User"
Enter any name you’d like in the "User Name" field. eg: docker-multi-travis-ci
Tick the "Programmatic Access" checkbox
Click "Next:Permissions"
Click "Attach Existing Policies Directly"
Search for "beanstalk"
Tick the box next to "AdministratorAccess-AWSElasticBeanstalk"
Click "Next:Tags"
Click "Next:Review"
Click "Create user"
Copy and / or download the Access Key ID and Secret Access Key to use in the Travis Variable Setup.

AWS Keys in Travis¶

Go to your Travis Dashboard and find the project repository for the application we are working on.
On the repository page, click "More Options" and then "Settings"
Create an AWS_ACCESS_KEY variable and paste your IAM access key
Create an AWS_SECRET_KEY variable and paste your IAM secret key

Deploying App¶

Make a small change to your src/App.js file in the greeting text.
In the project root, in your terminal run:

git add.git commit -m “testing deployment"git push origin main
Go to your Travis Dashboard and check the status of your build.
The status should eventually return with a green checkmark and show "build passing"
Go to your AWS Elasticbeanstalk application
It should say "Elastic Beanstalk is updating your environment"
It should eventually show a green checkmark under "Health". You will now be able to access your application at the external URL provided under the environment name.

Keys	Action
`?`	Open this help
`n`	Next page
`p`	Previous page
`s`	Search