Docker Up and Running Introduction

Docker Up and Running Introduction

• 1.
  Docker: Up andRunning Mark Beacom
• 2.
  Introduction Mark Beacom ● WithDrund since July 2014 ● Recently moved to DevOps from Backend ● I like coffee ● Rather be camping
• 3.
  What I’ll cover... ●What’s Docker? ● Benefits of Docker ● Images and Containers? ● The Docker ecosystem ● Dockerizing a flask web app ● Honorable mentions ● Live streaming: A resource intensive animal
• 4.
  What’s Docker? ● “Dockeris the world’s leading software container platform.” ● Established in 2008 ● Current release: v1.13 ● Github: 43K+ Stars (now moby/moby) ● Software platform allowing for rapid build, test, and deploy of applications ● Packages software into standardized units called containers ● Runs images of containers ● Wraps everything needed to run: code, runtime, system tools, system libraries - anything that can be installed on a server
• 6.
  Benefits of Docker ●Avoid: “Works on my machine” problems when collaborating on code ● Ship Software… Faster (7x) ● Improved Developer Productivity ● Easy deployment, scaling, and rollbacks ● Avoid: Application configuration drift ● Reusability!
• 7.
  Images... ● Lightweight ● Stand-alone ●Executable ● Immutable ● Created with: docker build ● Listed with: docker images ● Deployable to registries ● Generated from a Dockerfile definition The basis of a Docker container
• 8.
  Containers..? What’s acontainer? ● Runtime instance of an image ● The standard unit where the application or service resides ● Runs completely isolated from the host environment (unless configured) ● Runs natively on the host machine’s kernel ● Created with: docker run
• 9.
  Docker CLI ● Listingactive containers: docker ps ● Listing images: docker images ● Pulling images: docker pull alpine ○ Retrieves the alpine linux image from the Docker Hub repository ● Running an image: docker run alpine ls ○ Spins up a docker container running alpine with the ls command and returns the response.
• 10.
  Dockerize a Flaskwebapp! ● Collect the project source - We’re using Docker’s flask app example https://github.com/docker/labs/tree/master/beginner/flask-app ● Runs a python-based web application serving a single index template ● Defines all requirements, libraries, and configuration settings.
• 11.
  Anatomy of ourDockerfile ● FROM - Defines the base ● RUN - Executes commands during the build process ● COPY - Copies file(s) to the image ● EXPOSE - Exposes the port ● CMD - The default command executed at runtime
• 12.
  Initiating a Dockerbuild... docker build -t docker-demo . ● Results in a build with 9 steps ● Pulls the Alpine linux image from Docker Hub ● Runs our commands ● Copies assets ● Exposes the necessary port to the host
• 13.
  Building the image ●Step 1: FROM alpine:latest ● Step 2: RUN apk add --update py2-pip
• 14.
  Building the image ●Step 3: RUN pip install --upgrade pip ● Step 4: COPY requirements.txt /usr/src/app/
• 15.
  Building the image ●Step 5: RUN pip install --no-cache-dir -r /usr/src/app/requirements.txt
• 16.
  Building the image ●Step 6: COPY app.py /usr/src/app/ ● Step 7: COPY templates/index.html /usr/src/app/templates/ ● Step 8: EXPOSE 5000
• 17.
  Building the image ●Step 9: CMD python /usr/src/app/app.py
• 18.
  Running the container ●Mapping ports at runtime: ○ docker run -p 4000:5000 docker-demo ○ -p <host_port>:<container_port> ○ If no port is specified, a host port will be assigned randomly ● Naming the container for easier access and differentiation ○ docker run -p 4000:5000 --name demo docker-demo ○ --name <name> ○ If not specified, Docker will assign a random name to the container instance
• 19.
  Cleaning up... ● ListDocker containers: docker ps ● Stop a running container: docker stop <name_or_hash> ● Remove a stopped container: docker rm <name_or_hash> ● List images: docker images ● Remove an unused image: docker rmi <name_or_hash> A better way… ● docker run with --rm ● Automatically removes the container on exit
• 20.
  Live Demo Risky business…Let’s try it out!
• 21.
  Honorable Mentions ● DockerCompose ○ Defining a whole containerized workload in a single file definition ● Task Schedulers and Container Orchestration ○ Marathon, Amazon ECS, Rancher, Nomad
• 22.
  Our use casefor live streaming ● Deployed to Amazon Elastic Container Service ● Dynamically scales underlying resource pool based on utilization ● Individual transcoding containers independent for each stream ● Only use resources when the platform is in use ● Live stream capacity only limited by the maximum number of instances we can provision on AWS
• 23.
  Question and Answer