Once you get a bit of a rhythm with docker, it becomes pretty fun. Here are notes I took while reading some Docker books. Two good ones that I'd recommend are:
- Using Docker
- Docker in Action
docker inspect <containerid>- Shows IPAddress among other things.
docker diff <containerid>- List of files changed in the container.
docker logs <containerid>- list of everything that happened in the container.
docker start <containerid>- Start an exited/stopped container. START IS FOR CONTAINERS.
docker rm <containerid>- Remove a container.
docker run -it --name mynewcontainer centos bash- Run bash in a container called 'mynewcontainer' with centos as the base.
docker commit mynewcontainer arun/mynewcontainer- Create an image out of a container (either running or stopped).
docker run arun/mynewcontainer echo hi- Run echo on a new container and exit.
docker build -t arun/mynewcontainer .- Read Dockerfile from PWD and build a new container with that tag.
docker push arun/mynewcontainer- Push to docker hub or some other registry.
docker pull arun/mynewcontainer- Pull from docker hub or some other registry.
docker run arun/mynewcontainer–name test -d /usr/bin/nginx - daemonize and run a pulled image. RUN IS FOR IMAGES.
docker run --rm -it --link myredis:redis redis /bin/bash- Woah. We ran another container and linked it to the existing 'myredis' container as 'redis' on the new one, i.e. /etc/hosts has an entry called 'redis' pointing to the old one.
Old and busted
docker rmi $(docker images --filter "dangling=true" -q --no-trunc)- For images
docker rm -v $(docker ps -aq -f status=exited)- Remove all exited containers.
docker system prune
docker container prune
docker network prune
docker volume prune
Difference between Stopped Container and Image: stopped container retains its changes, settings, metadata, filesystem, runtime configuration, etc. Images don't have runtime information.
FROM centos:7 MAINTAINER Beech Team <email@example.com> RUN yum update && yum install -y emacs
Note: each command creates a layer, so try to squeeze stuff into a single command.
Then build it like this:
docker build -t test/myemacs .
And run it like this:
docker run test/emacs emacs .bashrc
If you add this line to the Dockerfile:
then you can directly run the container and pass the args without mentioning the command, i.e.
docker run test/emacs .bashrc
COPY in a Dockerfile instructs docker to copy a file from the host to the container. e.g.
COPY entrypoint.sh /
VOLUME is used to just use the host's filesystem for persistent data. e.g.
To do it at run time, do this:
docker run -it -v `pwd`:/root arunsrin/testpy
Instead of VOLUME, if you use ADD or COPY, it'll be baked into the image and available to anyone who downloads it. e.g. requirements.txt. Use ADD or COPY for making it part of the image, and VOLUME for sharing data between host and container. Unlike COPY, ADD also accepts URLs as a source, and unpacks it if its an archive. e.g.
ADD . /python-oauth
When running a command like
docker build -t asda/asd . , the '.' is
the build context. It's contents are tar'd and sent to the docker
daemon so that ADD and COPY commands work seamlessly. Don't run this
on a large folder like
docker-compose: stop all containers when one dies
docker-compose up --abort-on-container-exit
This will stop all containers launched via
docker-compose when one of the
Run redis on windows
My teams' integration tests need this and I'm the only windows guy!
docker run -d -p 6379:6379 redis:6