7.5 KiB
DOCKER BASICS
Table of Contents
INTRODUCTION
This file contains notes and code examples for using Docker.
I am a big fan of Academind on uDemy and YouTube. Their classes on eDemy are content rich and take you from noob to expert. Maximillian works from hands-on examples that you will actually use in IT. I highly recommend purchasing his course on Docker https://www.udemy.com/course/docker-kubernetes-the-practical-guide/.
UBUNTU INSTALLATION
Base Install
As usual, update your packages
sudo apt update
Prepare Ubuntu for the Docker installation
sudo apt install apt-transport-https ca-certificates curl software-properties-common
curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /usr/share/keyrings/docker-archive-keyring.gpg
echo "deb [arch=$(dpkg --print-architecture) signed-by=/usr/share/keyrings/docker-archive-keyring.gpg] https://download.docker.com/linux/ubuntu $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
Update the packages again
sudo apt update
Now do the install from Docker
apt-cache policy docker-ce
sudo apt install docker-ce
Make sure it is running
sudo systemctl status docker
Put yourself in the Docker group
Do this so you don't have to use sudo all the time
sudo usermod -aG docker ${USER}
You can reboot to make it take effect or...
su - ${USER}
CONCEPTS
Images
Images are the template (instructions) that define the parameters of the container. The image is used to build containers and a single image can build many containers.
Containers
Containers are packages that contain the application and the environment in which the application runs.
IMAGES
Define your own image using a Dockerfile
Create a Dockerfile in the project directory. This file is a set of instructions for building a container.
# the baseImage is either a name from DockerHub or is the
# name of an Image on your system
FROM node
# tell Docker where to run commands from
WORKDIR /app
# tell Docker which files go into the image
# COPY localFolder imageFolder
COPY . /app
# do some node stuff
RUN npm install
# show the port to the world outside of Docker so that the
# web pages are visible
EXPOSE 80
# these commands are executed when a container is instanced
CMD ["node", "server.js"]
Build an image from the Dockerfile.
Using this command results in a random Image ID that you use to run a container. This is not the way.
docker build .
Assign a Name:Tag to the image
Assign a name and tag to the image because you are a responsible dev.
docker build -t name:tag .
MANAGING IMAGES
List out the images installed on this server
docker images
Remove an image from this server
docker rmi {imagename}
Remove all dangling (untagged) images
docker image prune
CONTAINERS
Run a container from a DockerHub image
docker run node
This command will create and run a container instance from the image. In this example it is using the office nodejs image to create a running instance in a container. Note: this command as it is doesn't do much.
docker run -it node
Adding the -it flag creates an interactive terminal instance where you can interact with the instance of node.
Running a container from your own image
After the image is built you will get an ID for the image. Use the ID to start the container. If you are a good dev, then you assigned a name:tag and you use that instead.
docker run -p 3000:80 e5a737911cca
MANAGING CONTAINERS
Parameters for running a container
| Switch | Description |
|---|---|
| -p | (port) map a port |
| -d | (detached mode) run in the background |
| -it | (interactive mode) not -d |
| --rm | (remove) the container when the service is stopped |
| --name | (name) give the container a name |
| -v | (volume) assigns a named volume to the container |
docker run -p 3000:80 -d --rm --name goalsapp goals:latest
List all running containers
docker ps
List all containers (running and not running)
docker ps -a
Stop a container
docker stop {containername}
Remove a stopped container
docker rm {containername}
Remove ALL stopped containers
docker container prune
Volumes
Volumes store persistent data that survives after a container shuts down. These are folders on the host machine that are mounted in the container.
Volumes are managed by Docker. This means that we do not know where the volume is located and have no ability to manage them. Bind mounts solve that problem.
Add a volume to the container
This command creates an anonymous volume. Anonymous volumes are not persisted when the container is removed.
VOLUME ["{path}"]
VOLUME ["app/feedback"]
Named volumes persist data when the container is removed. Named volumes are NOT defined in the Dockerfile, rather it is done at the command line when running the container.
-v {volume name}:{mount location in container}
docker run -d -p 3000:80 --rm --name feedback-app -v feedback:/app/feedback feedback-node:volumes
List volumes
docker volume ls
Troubleshooting
docker logs {appname}
Bind Mounts
Like named volumes but we have management control over the file system. Bind mounts are not created in the Dockerfile. Bind mounts are created from terminal.
A bind mount is created by adding a second -v command. Should use quotes to protect the absolute path.
-v "{absolute path on host file system}:{mount location in container}"
docker run -d -p 3000:80 --rm --name feedback-app -v feedback:/app/feedback -v "/home/kschaefer/repos/webdev/docker/data-volumes-01-starting-setup:/app" feedback-node:volumes
If you mount a volume to the app folder in the way that this command above does, then the mount will overwrite everything that the COPY and RUN commands do in the Dockerfile. This command is overwriting the container folder with the contents of the local host folder. Since the local host does not have the same npm dependencies as the container then we get an issue.
The solution is to use an anonymous volume. In the following command the folder that is defined in the anon volume is protected.
docker run -d -p 3000:80 --rm --name feedback-app -v feedback:/app/feedback -v "/home/kschaefer/repos/webdev/docker/data-volumes-01-starting-setup:/app" -v /app/node_modules feedback-node:volumes
🛸 Why do this? Now we can make changes to the HTML files and we do NOT have to rebuild the container.
Read Only Volume
When using a Bind Mount, the local drive that is mapped to the container is writable by the container. Since we probably don't want the container writing back to our source-code marking the volume as read-only is a good idea. Do this by adding :ro to the end of the volume declaration.
docker run -d -p 3000:80 --rm --name feedback-app -v feedback:/app/feedback -v "/home/kschaefer/repos/webdev/docker/data-volumes-01-starting-setup:/app:ro" -v /app/node_modules feedback-node:volumes
You can add volume declarations after the bind-mount to make subfolders writable if they need it. Use anonymous folders to do this. In the above example, all of app is read only but node_modules can be written by the container.