Docker

Tool for running app in isolated environment

Container

Abstraction at the app layer that packages code and dependencies together. Multiple containers can run on SAME machine, each as an ISOLATED process

• A running instance of an Image

Virtual Machines

Abstraction of physical hardware-->turn one server into many. Full copy of OS, app, and libraries; Takes up huge space and slow to boot

Docker Image

• Template for creating an environment

• Snapshot

• Has everything need to run app (OS, Software, App code)

General Process

1. Pull official Images from Docker Hub: docker pull nginx and check all Images: docker images

2. Run a container in Image: docker run -d -p 8080:80 nginx:latest; "latest" is the TAG and needed

3. Check all containers running: docker container ls

4. Remove container: docker rm [ID/NAME] / docker rm $(docker ps -aq) / docker rm -f $(docker ps -aq)

Exposing Port

• The container is running with a port of 80

• Host issue request of localhost:8080 mapped to PORT 80 in container --> run -p 8080:80 when running the container

• Open up browser and go to localhost:8080 to get the container's application's page

• Multiple hosts to one port: -p 8080:80 -p 3000:80

Naming Container

• docker run --name website -d -p 3000:80 -p 8080:80 nginx:latest Specify the name, easier to identify and stop

Volumes

Allow sharing of data, files, and folders

• Hosts and container form Volumes

• Create file locally: website/index.html (website is a folder)

• docker run --name website -v $(pwd):/usr/share/nginx/html:ro -d -p 8080:80 nginx:latest. In this way, we serve the container with our local file (ro Read Only flag can be removed in order to write into container)

• docker exec -it website bash. In this way, we will be inside the container. Go to usr/share/nginx/html to make changes, and these changes will be on local machine as well

Share Volumes Between Containers

• docker run --name website-copy --volumes-from website -d -p 8081:80 nginx:latest. In this way, we use the data from website container to serve the website-copy container.

DockerFile

Create our own images

Dockerfile referenceDocker Documentation

• Create a file called "Dockerfile" in the root directory of project

FROM nginx:latest
ADD . /usr/share/nginx/html

• docker build --tag website:latest . (. stands for current directory, this is where to look for a Dockerfile)

• Can use node.js and express framework to make a simple backend server

FROM node:latest
WORKDIR /app
ADD . .
RUN npm install
CMD node index.js

• Go into your local node.js app folder: docker build --tag user-api:latest .

• docker run --name user-api -d -p 3000:3000 user-api:latest This runs the container

.dockerignore

node_modules
Dockerfile
.git
folder/**

Layer and Cache

Use cache to boost performance: A source file change shouldn't trigger all things to rebuild

FROM node:latest
WORKDIR /app
ADD package*.json ./ 
RUN npm install
ADD . .
CMD node index.js

In this way, adding package.json & package-lock.json into the container is using cache since these two files are not changing.

Reduce Image Size

Use ALPINE; Supported tags for official images

docker pull node:lts-alpine --> smaller size and faster

Tags, Versioning, and Tagging:

• Allow control image version

• Avoid breaking changes (Not always using the latest version)

• Safe

Have control over the tags + version: docker pull node:8-alpine instead of docker pull node:alpine

How to correctly Tag images

docker tag website:latest website:1 Tag from latest to 1, so that tag 1 contains current version. If do another docker tag website:latest website:2, tag 2 will be pointing to the latest one, and changes in the latest version will not reflect in tag 1

Docker Registries

• Server application that can store and distribute Docker images. Store images in a server host instead of our local host

• Used in CD/CI Pipeline

• Running our applications by uploading container from local to registry.

Host have images --> Have a docker registry --> push images from host to docker registry

Private/Public Registry

Create Docker Hub Repo

1. Create Repo from Docker Hub website (Similar to Github)

2. docker push user/website:tagname to push new tag to the repo

Inspect and Logs

docker logs [containerID] to show all logs

docker logs -f [containerID] to follow the log and see real-time updates when new logs happen

docker inspect [containerID] to inspect detailed information of this container