Lab > Dockerizing applications

Using Docker to contain the application's code, and all the necessary files to run on a server, locally, or in a Kubernetes swarm.

• Objective and Tasks

Why Docker?

We used to have virtual machines that encapsulated our applications and logic, and at the height of the virtual machines sometimes we almost had a virtual machine for every task it seemed. When the virtual machines started to become cumbersome and bloated, containers were conceptualized with the philosophy of containing all the relevant pieces to run on top of a specialized bare metal system; meaning when the shared pieces of the virtual machines were abstracted away into Docker; the containers would run inside these specialized virtual machines. In essence the containers a something like a highly specialized .zip.

Containers, and The Registry

Containers themselves are not useful without using the specialized .zip file masquerading as a container. Success is had with a container when it runs on your machine. When the container runs on your machine and you're ready to publish it out to the world; there is a public and various private registries that allow sharing the containers across the ecosystem. The registry provides a catalog of containers with their relevant configuration.

Building containers from containers

Another feature of containers is the ability to start from an existing ecosystem like node, python, golang, etc. and leverage the existing containers as a starting point or to create a custom one that for what you need it for. When containers stack upon each other they take with them the good and bad parts of the container. In Docker's infancy, the containers were pretty large--to alleviate some of the container bloat pressures there were containers like the alpine:node which specializes in only having the relevant pieces needed to run the application.

This is an example of the container stacks that derive from building various containers on top of each other.

Dockerfile(s)

Here is a brief sample of a Dockerfile that will build a container that will run a simple node application.

FROM node:latest

WORKDIR /app

COPY . .

RUN npm install

RUN npm run build

CMD ["npm", "start"]

Building the container

docker build --tag dockerized-app .

Running Containers

Here is the documentation on running a node like application.

docker run --publish 3000:3000 dockerized-app

Running in detached mode

This is great so far, but our sample application is a web server and we should not have to have our terminal connected to the container. Docker can run your container in detached mode or in the background. To do this, we can use the --detach or -d for short. Docker will start your container the same as before but this time will “detach” from the container and return you to the terminal prompt.

docker run --detach --publish 3000:3000 dockerized-app

docker pull/push

Like git docker has commands that mimic git's push/pull philosophy. Docker can leverage the similar concepts and the parrallel is the same as git's, you pull down changes. When you've changed the container, and it's code--you' ll docker build, and docker push out the containers to the relevant registries.