Docker: The basics

by Joe Meilinger

What/why docker?

  • Lightweight "Virtual Machine" runner
  • Enforces much of 12-factor application design
  • Provides a simpler mechanism to move an application from development to other environments
  • Reduces ops burden when utilizing a container orchestrator (eg. Kubernetes, Kubernetes derivates, docker swarm, etc) versus managing application servers

Text

Virtual Machines vs Docker Containers

Images vs containers

  • Images are built by running commands contained in a Dockerfile
  • Images are stored in a registry and contain a complete binary representation of the layers used to create the image
  • Containers are running images
  • In OOP: Image = Class
  • In OOP: Container = Object
  • Containers can have volumes mapped, run in a network, have environment variables injected, etc
  • Containers are meant to be lightweight and disposable
  • Runs a single process (which can spawn others as needed)

Getting setup

  • Install Docker for Windows/Docker for Mac
  • From bash/git-bash, verify docker is running properly via:
> docker ps  # list running containers

CONTAINER ID      IMAGE       COMMAND        CREATED      STATUS      PORTS

> docker images  # list images in local registry

REPOSITORY        TAG      IMAGE ID       CREATED      SIZE

Basic Docker cli commands

  • docker ps -- displays list of running containers
  • docker images -- displays a list of local registry images
  • docker build -- build a docker image
  • docker run -- run a container for an image
  • docker exec -- run a process on a running container
  • docker logs -- display STDOUT/STDERR of container
  • docker system prune -- clean up your local system

Building an image

  • Uses a Dockerfile to build
  • By default, publishes image to a local registry
  • Never build actual configuration into images, defaults are fine
  • Same image should be able to be used across environments as configuration would be the only thing changing
> cat Dockerfile

FROM python:3.8
WORKDIR /app
COPY . /app
RUN pip install -r requirements.txt
ENTRYPOINT ["python", "app.py"]

> docker build .  # builds docker image using auto-generated name and ID

> docker build . -t my-image  # builds docker image, tagging w/ "latest"

> docker build . -t my-image:1.0  # builds docker image, tagging w/ "1.0"

Running a container from an image

  • Pulls image from a registry (local or remote)
  • Runs on a docker host (local machine or remote host)
> docker run [IMAGE]  # runs image by repo:tag OR image id using default entrypoint

> docker run -it [IMAGE] /bin/bash  # runs image using your specified entrypoint (/bin/bash)

> docker run -p 3000:5000 [IMAGE]  # runs image, mapping port 5000 on the container 3000 on host

Image Registry

img1

img2

img3

container1

from img1

Docker runtime

Mapping volumes/attaching to containers/injecting environment variables etc

> docker run -v "$(pwd)":/app [IMAGE]  # run image, mount current host folder to /app on container

> docker exec -it [CONTAINER_ID] /bin/bash  # execute command on running container (/bin/bash)

> cat vars

VAR1=VAL1
VAR2=VAL2

> docker run --env-file vars [IMAGE]  # runs image, using env variables from "vars" file
  • Able to overlay a host filesystem location to a container's filesystem via bind mount
  • Can run an additional process (ex. interactive shell) on a running container via "docker exec" command

Networking containers together

  • Docker utilizes internal/bridged networks
  • If you're running more than one container, probably easiest to create a docker-compose.yml to orchestrate a stack

API

UI

Docker runtime

FGDRT + C3PO example

  • Run local C3PO + API via docker-compose
  • Run FGDRT API + FGDRT UI

Docker runtime

C3PO DB

C3PO API

FGDRT API

FGDRT UI

localhost:3001

> docker-compose -d --build
> docker-compose --build

Resources

Step-by-step tutorial/example

Made with Slides.com