Docker Cheat Sheet - 2021 Edition

Docker Cheat Sheet

Docker Cheat Sheet - 2021 Edition

Docker delivers software in containers, which simplifies the process by packaging everything it takes to run an application. With Docker growing in popularity, it becomes necessary to understand it before using it on your own or for a business. After all, PayPal, Netflix, AT&T, Oracle, Microsoft, and other major businesses use Docker as their preferred tool. That’s why we created this Docker cheat sheet for you. It will be of use before a job interview or simply to refresh your memory on certain commands. You might come upon to help assist you in learning by yourself, which would be even more detailed.

Start your 30-day FREE TRIAL to learn more about our Docker courses here at CloudInstitute.io.

Check Version

It’s good to know the current version of Docker. This will help you know what features are compatible with what you currently running. It’s also useful to know what containers to run whenever you are getting template containers. Here’s how to figure out what version you’re running:

Find server version:

$ docker version --format '[[.Server.Version]]' 1.8.0

Here’s how to dump raw JSON data:

$ docker version --format '[[json .]]' {"Client":{"Version":"1.8.0","ApiVersion":"1.20","GitCommit":"f5bae0a","GoVersion":"go1.4.2","Os":"linux","Arch":"am"}

Containers

Docker is all about containers. It’s simply a package of software that can include code, tools, libraries, settings and anything else you need. It’s all in one place, one container.

Lifecycle

If you run a container without options, it will start and stop instantaneously. To keep it running, use the command, docker run -td container_id 

The option -t will assign a pseudo-TTY session and -d will detach the container (run container in background and print container ID). To have a temporary container, docker run --rm  removes the container after it stops

To delete volumes associated with the container, deleting it should include the -v switch such as in docker rm -v.

To run docker with a custom log driver (i.e., to syslog), use docker run --log-driver=syslog. Another option is docker run --name yourname docker_image because when you specify the –name inside the run command this will allow you to start and stop a container by naming it what you specified.

Starting and Stopping

If you want to detach from a running container, use Ctrl + p, Ctrl + q. To combine a container with a host process manager, start daemon with -r=false then use docker start -a.

CPU Constraints

Limit CPU by using a percentage of all CPUs or by using specific cores.

For instance, you can tell the cpu-shares setting. The setting is odd. 1024 means 100% of the CPU. To have the container take 50% of the CPU cores, you specify 512.

docker run -it -c 512 agileek/cpuset-test

Another thing you can do is use CPU cores with cpuset-cpus.

Memory Constraints

Setting memory constraints on Docker:

docker run -it -m 300M ubuntu:14.04 /bin/bash

Capabilities

Linux capabilities are set by using cap-add and cap-drop for better security.

In order to mount a FUSE based filesystem, you need to combine both --cap-add and --device:

docker run --rm -it --cap-add SYS_ADMIN --device /dev/fuse sshfs

Give access to a single device:

docker run -it --device=/dev/ttyUSB0 debian bash

Give access to all devices:

docker run -it --privileged -v /dev/bus/usb:/dev/bus/usb debian bash

Info

docker ps -a shows running as well as stopped containers.

docker stats --all shows a list of all containers, default shows the running ones.

Import / Export

  • docker cp copies files/folders between a container and local filesystem.
  • docker export turns container filesystem into tarball archive stream to STDOUT.

Executing Commands

Enter a running container by attaching a new shell process to a running container. This is called foo, and use: docker exec -it foo /bin/bash.

Images

Images are templates for Docker containers.

Lifecycle

  • docker images show images.
  • docker import creates an image from tarball.
  • docker build creates an image from Dockerfile.
  • docker commit creates an image from a container, pausing it temporarily if it’s running.
  • docker rmi removes an image.
  • docker load loads an image from a tar archive as STDIN, including images and tags.
  • docker save saves images to a tar archive stream to STDOUT with all parent layers, tags and versions.

Info

Cleaning up

You can always use the docker rmi command to remove images, but a tool called docker-gc can safely clean up images that aren’t used by any containers. As of docker 1.13, docker image prune can also remove unused images.

Load/Save Image

Load an image from file:

docker load < my_image.tar.gz

Save an existing image:

docker save my_image:my_tag | gzip > my_image.tar.gz

Import/Export Container

Import a container (as an image from file):

cat my_container.tar.gz | docker import - my_image:my_tag

Export an existing container:

docker export my_container | gzip > my_container.tar.gz

Difference Between Loading Saved Image and Importing an Exported Container as an Image

Load command will create a new image, and this will include the history.
Importing a container as an image using the import command creates a new image, but not the history, which creates a smaller image size than loading an image.

Networks

Docker automatically creates three network interfaces when you install it—bridge, host and none. By default, a new container is launched into the bridge network. To enable communication between multiple containers, create a new network and launch containers in it. This enables containers to communicate to one another while being apart from containers that are not connected to the network. It also helps map container names to their IP addresses.

Lifecycle

  • docker network create NAME Create a new network (default type: bridge).
  • docker network rm NAME Remove one or more networks by name or identifier. Containers can’t be linked to the network when you’re deleting them.

Info

Connection

Here is how you can assign a unique IP address for a container:

# create a new bridge network with your subnet and gateway for your ip blockdocker network create --subnet 203.0.113.0/24 --gateway 203.0.113.254 iptastic # run a nginx container with a specific ip in that block$ docker run --rm -it --net iptastic --ip 203.0.113.2 nginx # curl the ip from any other place (assuming this is a public ip block duh)$ curl 203.0.113.2

Registry & Repository

A collection of tagged images is a repository, and it’s hosted.

A registry is a host. This means it’s a server that stores repositories and provides an HTTP API for managing, uploading and downloading repositories.

Docker.com includes a massive number of repositories. However, images on Docker may not be secure, as there are some unverified images on its repository.

Dockerfile

Dockerfile creates a Docker container when you run docker build on it. This is preferred over docker commit.

Instructions

  • .dockerignore
  • FROM Sets the Base Image for resulting instructions.
  • MAINTAINER (deprecated - use LABEL instead) Set the Author field of the generated images.
  • RUN execute any commands in a new layer on top of the current image and commit the results.
  • CMD provide defaults for an executing container.
  • EXPOSE informs Docker that the container listens on the specified network ports at runtime. This doesn’t make ports accessible.
  • ENV sets environment variable.
  • ADD copies new files, directories or remote file to container. Invalidates caches. Avoid ADD and use COPYinstead.
  • COPY copies new files or directories to container. The default is to copy as a root despite any of the USER/WORKDIR settings. Use --chown=<user>:<group> to give ownership to another user/group. (Same for ADD.)
  • ENTRYPOINT configures a container that will run as an executable.
  • VOLUME creates a mount point for externally mounted volumes or other containers.
  • USER sets the user name for following RUN / CMD / ENTRYPOINT commands.
  • WORKDIR sets the working directory.
  • ARG defines a build-time variable.
  • ONBUILD adds a trigger instruction when the image is used as the base for another build.
  • STOPSIGNAL sets the system call signal that will be sent to the container to exit.
  • LABEL apply key/value metadata to your images, containers, or daemons.
  • SHELL override default shell is used by docker to run commands.
  • HEALTHCHECK tells docker how to test a container to check that it is still working.

[[widget type="Itskills\BlogWidget\Block\Widget\Category" title="Enroll in Our Docker Certifications Training Program" description="Cloud Institute offers Docker certifications training that has helped thousands of learners prepare for a rewarding career." button_title="Get Started" id_path="category/1787"]]

Layers

The filesystem in Docker is based on layers.

Links

Links are how Docker containers communicate with one another through TCP/IP ports. 

Volumes

It’s not necessary to have Docker volumes connected to any container. You can always use volumes attached from data-only containers for portability. Docker has names volumes which replace data-only containers. It’s best to use named volumes to implement it rather than data containers.

Lifecycle

Info

Volumes are helpful to use when you can't use links (which are TCP/IP only). For an example, if you want to have two docker instances communicate by leaving things on the filesystem.

One solution is to stack them in several docker containers at once, and this is through using docker run --volumes-from.

Volumes are isolated, so they’re usually used to store state from divisions between temporary containers.

Here’s how to map MacOS host directories as volumes:

docker run -v /Users/wsargent/myapp/src:/src

You can also run data-only containers as described to provide data portability.

Finally, you can also mount files as volumes.

Exposing Ports

Exposing incoming ports through the host container is completed by mapping the container port to the host port (only using localhost interface) using -p:

docker run -p 127.0.0.1:$HOSTPORT:$CONTAINERPORT --name CONTAINER -t someimage

You can inform Docker how a container listens on the network ports at runtime by using command EXPOSE:

EXPOSE <CONTAINERPORT>

EXPOSE doesn’t expose the port itself -- only -p will do that. Here’s how you can expose the container's port:

iptables -t nat -A DOCKER -p tcp --dport <LOCALHOSTPORT> -j DNAT --to-destination <CONTAINERIP>:<PORT>

Running Docker in Virtualbox? Then you’ll have to forward the port there too, using forwarded_port. You’ll need to outline a port range in your Vagrantfile like this so you can map them dynamically:

Vagrant.configure(VAGRANTFILE_API_VERSION) do |config|  ...   (49000..49900).each do |port|    config.vm.network :forwarded_port, :host => port, :guest => port  end   ...end

If you’ve forgotten what you mapped the port to on the host container, use docker port to show it:

docker port CONTAINER $CONTAINERPORT

Docker-Compose

Compose defines and runs multi-container Docker applications. Compose is how you can use a YAML file to configure your app’s services. A single command can be used to create and begin all services from your configuration.

You can use the following command to start your application:

docker-compose -f <docker-compose-file> up

You can also run docker-compose in detached mode using -d flag, which will help you stop whenever needed by the following command:

docker-compose stop

You can remove containers with the down command. Pass --volumes to also remove the data volume.

Security

Docker runs as root. If you’re inside the docker group, you’ll have root access. If you expose the docker unix socket to a container, you’re providing the container root access to the host. Docker shouldn’t be left defenseless, so you should still make sure the security is solid.

Security Tips

For the best security, you want to run Docker inside a virtual machine. This is the recommendation from the Docker Security Team Lead.

Sensitive information, image ids, shouldn’t be given to anyone. It’s best to think of them as passwords and not to give them out.

Since docker 1.11 came out, you can limit the number of active processes running inside a container to prevent fork bombs. This needs a linux kernel >= 4.3 with CGROUP_PIDS=y to be in kernel configuration.

docker run --pids-limit=64

Another addition since docker 1.11 is the capability to prevent processes from gaining new privileges. This feature has been in the linux kernel since version 3.5.

docker run --security-opt=no-new-privileges

Here are Container Solutions:

Turn off interprocess communication with:

docker -d --icc=false --iptables

Set the container to be read-only:

docker run --read-only

Verify images with a hashsum:

docker pull debian@sha256:a25306f3850e1bd44541976aa7b5fd0a29be

Set volumes to be read only:

docker run -v $(pwd)/secrets:/secrets:ro debian

Define/run a user in your Dockerfile. This will help you not run as root inside the container:

RUN groupadd -r user && useradd -r -g user userUSER user

Security Roadmap

The Docker roadmap talks about seccomp support. Bane, an AppArmor policy generator is useful for a security roadmap, and they're working on security profiles.

Tips

Some tips:

Prune

 Data Management Commands are available as of Docker 1.13:

  • docker system prune
  • docker volume prune
  • docker network prune
  • docker container prune
  • docker image prune

df

docker system df is a basic summary of space currently used by docker objects.

Heredoc Docker Container

docker build -t htop - << EOFFROM alpineRUN apk --no-cache add htopEOF

Last Ids

alias dl='docker ps -l -q'docker run ubuntu echo hello worlddocker commit $(dl) helloworld

Commit with Command (Requires Dockerfile)

docker commit -run='{"Cmd":["postgres", "-too -many -opts"]}' $(dl) postgres

Get IP Address

docker inspect $(dl) | grep -wm1 IPAddress | cut -d '"' -f 4

or with jq installed:

docker inspect $(dl) | jq -r '.[0].NetworkSettings.IPAddress'

or using a go template:

docker inspect -f '[[ .NetworkSettings.IPAddress ]]' <container_name>

or when building an image from Dockerfile, when you want to pass in a build argument:

DOCKER_HOST_IP=`ifconfig | grep -E "([0-9]{1,3}\.){3}[0-9]{1,3}" | grep -v 127.0.0.1 | awk '{ print $2 }' | cut -f2 -d: | head -n1`echo DOCKER_HOST_IP = $DOCKER_HOST_IPdocker build \  --build-arg ARTIFACTORY_ADDRESS=$DOCKER_HOST_IP   -t sometag \  some-directory/

Get port mapping

docker inspect -f '[[range $p, $conf := .NetworkSettings.Ports]] [[$p]] -> [[(index $conf 0).HostPort]] [[end]]' <containername>

Find containers by regular expression

for i in $(docker ps -a | grep "REGEXP_PATTERN" | cut -f1 -d" "); do echo $i; done

Get Environment Settings

docker run --rm ubuntu env

Kill running containers

docker kill $(docker ps -q)

Delete all containers (force!! running or stopped containers)

docker rm -f $(docker ps -qa)

Delete old containers

docker ps -a | grep 'weeks ago' | awk '{print $1}' | xargs docker rm

Delete stopped containers

docker rm -v $(docker ps -a -q -f status=exited)

Delete containers after stopping

docker stop $(docker ps -aq) && docker rm -v $(docker ps -aq)

Delete dangling images

docker rmi $(docker images -q -f dangling=true)

Delete all images

docker rmi $(docker images -q)

Delete Dangling Volumes

As of Docker 1.9:

docker volume rm $(docker volume ls -q -f dangling=true)

In 1.9.0, the filter dangling=false does not work - it is ignored and will list all volumes.

Show Image Dependencies

docker images -viz | dot -Tpng -o docker.png

Slimming Down Docker Containers

  • Cleaning APT in a RUN layer

This should be done in the same layer as other apt commands. Otherwise, the previous layers still persist the original information and your images will still be fat.

RUN {apt commands} \  && apt-get clean \  && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*

  • Flatten an image

ID=$(docker run -d image-name /bin/bash)docker export $ID | docker import – flat-image-name

  • For backup

ID=$(docker run -d image-name /bin/bash)(docker export $ID | gzip -c > image.tgz)gzip -dc image.tgz | docker import - flat-image-name

Monitor System Resource Utilization for Running Containers

To check the CPU, memory, and network I/O usage of a single container, you can use:

docker stats <container>

For all containers listed by id:

docker stats $(docker ps -q)

For all containers listed by name:

docker stats $(docker ps --format '[[.Names]]')

For all containers listed by image:

docker ps -a -f ancestor=ubuntu

Remove all untagged images:

docker rmi $(docker images | grep “^” | awk '{split($0,a," "); print a[3]}')

Remove container by a regular expression:

docker ps -a | grep wildfly | awk '{print $1}' | xargs docker rm -f

Remove all exited containers:

docker rm -f $(docker ps -a | grep Exit | awk '{ print $1 }')

Volumes Can Be Files

Be aware that you can mount files as volumes. For example you can inject a configuration file like this:

# copy file from containerdocker run --rm httpd cat /usr/local/apache2/conf/httpd.conf > httpd.conf # edit filevim httpd.conf # start container with modified configurationdocker run --rm -it -v "$PWD/httpd.conf:/usr/local/apache2/conf/httpd.conf:ro" -p "80:80" httpd

Docker Training

We hope you found this short cheat sheet useful. We can provide even more explanations, from the basics of Docker all the way up to advanced concepts and practical training. Start your 30-day FREE TRIAL  to gain access to over 200 self-paced courses. 

Talk to our experts to learn more.

Previous Post Next Post
Hit button to validate captcha