Docker cheatsheet¶

Compiled from different projects.

Table of Contents¶

Command line
Recipes
1. Clustering
2. Commiting Dockerfile Modifications

Command line¶

The docker CLI provides a wide range of methods for creating, maintaining, and removing both docker images and containers.

Running images as containers¶

The docker run command offers many flags to change the behaviour of a docker container:

-d : detach; run the container using the daemon (i.e. not in the current shell)
-p : port; bind some -p HOST:CONTAINER port to the container
-v : volume; share some directory -v HOST:CONTAINER with the container

You can pass environment variables using the -e flag, followed by a pair-value string

-e "MY_ENV_VAR=VALUE"

Distributing docker containers¶

After a image has been built, it may be exported with

docker save -o [image-name].tar [image-name]

and loaded (e.g. on another machine)

docker load -i [image-name].tar

You can also commit a running container, and save the state of any modifications you have made into a new image with

docker commit [container-id/name]  new-image-name

All of the docker images can be listed with

docker image ls 
# or
docker images

Viewing container status¶

A simple was of viewing the container status us to view the log files it produces (commonly just STDOUT in the container)

docker logs [container-id/name]

with an optional -f flag for following the output.

Another method is to attach to a running process if it is running in the background and/or started with the -d flag

docker attach [container-id/name]

NB: using ctrl-C or ctrl-D will kill the container much as if it were running in the current shell; use ctrl-p or ctrl-q instead.

You can spawn an interactive shell on a running container using

docker exec -it [container-id/name] /bin/bash

You can list the containers currently running with

docker ps

or all running and exited containers with

docker ps -a

A JSON structure containing all information on a running container can be found with

docker inspect [container-id/name]

Copying files¶

Files can be coied to and from running containers using

docker cp [local-path] [container-id/name]:[remote-path]

### Cleanup Following instructions from Hostinger, we can employ the following commands to clear unneeded containers, images, networks, and more:

docker image prune

removes recent, dangling and untagged images. This command will commonly have the -a flag to ensure all related files are deleted. You can also apply filters such as --filter "until=24h" to delete images created within a specific time frame.

To remove idle containers that have been stopped, use

docker container prune

For volumes and networks, the commands are likewise very similar

docker volume prune
docker network prune

To do all of these prune commands in one, execute

docker system prune

Recipes¶

A few case studies in using different docker commands.

Clustering¶

I recently set up a VerneMQTT cluster using docker images, which require knowledge of the IP addresses of the first host. This information can be found on a running container using the

docker inspect [container-id/name] | jq ".[0].NetworkSettings.IPAddress"

command, which pipes the JSON structure into JQ, and extracts specifically just the IP address.

Commiting Dockerfile Modifications¶

If you have a running container, but wish to modify it’s behaviour so that port 80 is always exposed, you can do so with the commit --change flag:

docker commit --change="EXPOSE 80" [container-id/name] new-image-name
# --change="" can also just be -c ""

notes

Docker cheatsheet

Contents