The dip is a CLI dev–tool that provides native-like interaction with a Dockerized application. It gives the feeling that you are working without using mind-blowing commands to run containers.

• Local development with Docker containers
• Dockerized Ruby on Rails application
• Dockerized Node.js application: one, two
• Dockerized Ruby gem
• Dockerizing Ruby and Rails development
• Reusable development containers with Docker Compose and Dip

gem install dip

Dip can be injected into the current shell (ZSH or Bash).

eval "$(dip console)"

IMPORTANT: Beware of possible collisions with local tools. One particular example is supporting both local and Docker frontend build tools, such as Yarn. If you want some developer to run yarn locally and other to use Docker for that, you should either avoid adding the yarn command to the dip.yml or avoid using the shell integration for hybrid development.

After that we can type commands without dip prefix. For example:

<run-command> *any-args
compose *any-compose-arg
up <service>
ktl *any-kubectl-arg
provision

When we change the current directory, all shell aliases will be automatically removed. But when we enter back into a directory with a dip.yml file, then shell aliases will be renewed.

Also, in shell mode Dip is trying to determine manually passed environment variables. For example:

VERSION=20180515103400 rails db:migrate:down

You could add this eval at the end of your ~/.zshrc, or ~/.bashrc, or ~/.bash_profile. After that, it will be automatically applied when you open your preferred terminal.

dip --help
dip SUBCOMMAND --help

The configuration is loaded from dip.yml file. It may be located in a working directory, or it will be found in the nearest parent directory up to the file system root. If nearby places dip.override.yml file, it will be merged into the main config.

Also, in some cases, you may want to change the default config path by providing an environment variable DIP_FILE.

Below is an example of a real config. Config file reference will be written soon. Also, you can check out examples at the top.

# Required minimum dip version
version: '8.0'

environment:
  COMPOSE_EXT: development
  STAGE: "staging"

compose:
  files:
    - docker/docker-compose.yml
    - docker/docker-compose.$COMPOSE_EXT.yml
    - docker/docker-compose.$DIP_OS.yml
  project_name: bear

kubectl:
  namespace: rocket-$STAGE

interaction:
  shell:
    description: Open the Bash shell in app's container
    service: app
    command: bash
    compose:
      run_options: [no-deps]

  bundle:
    description: Run Bundler commands
    service: app
    command: bundle

  rake:
    description: Run Rake commands
    service: app
    command: bundle exec rake

  rspec:
    description: Run Rspec commands
    service: app
    environment:
      RAILS_ENV: test
    command: bundle exec rspec

  rails:
    description: Run Rails commands
    service: app
    command: bundle exec rails
    subcommands:
      s:
        description: Run Rails server at http://localhost:3000
        service: web
        compose:
          run_options: [service-ports, use-aliases]

  stack:
    description: Run full stack (server, workers, etc.)
    runner: docker_compose
    compose:
      profiles: [web, workers]

  sidekiq:
    description: Run sidekiq in background
    service: worker
    compose:
      method: up
      run_options: [detach]

  psql:
    description: Run Postgres psql console
    service: app
    default_args: db_dev
    command: psql -h pg -U postgres

  k:
    description: Run commands in Kubernetes cluster
    pod: svc/rocket-app:app-container
    entrypoint: /env-entrypoint
    subcommands:
      bash:
        description: Get a shell to the running container
        command: /bin/bash
      rails:
        description: Run Rails commands
        command: bundle exec rails
      kafka-topics:
        description: Manage Kafka topics
        pod: svc/rocket-kafka
        command: kafka-topics.sh --zookeeper zookeeper:2181

  setup_key:
    description: Copy key
    service: app
    command: cp `pwd`/config/key.pem /root/keys/
    shell: false # you can disable shell interpolations on the host machine and send the command as is

  clean_cache:
    description: Delete cache files on the host machine
    command: rm -rf $(pwd)/tmp/cache/*

provision:
  - dip compose down --volumes
  - dip clean_cache
  - dip compose up -d pg redis
  - dip bash -c ./bin/setup

Current OS architecture (e.g. linux, darwin, freebsd, and so on). Sometime it may be useful to have one common docker-compose.yml and OS-dependent Compose configs.

Relative path from the current directory to the nearest directory where a Dip's config is found. It is useful when you need to mount a specific local directory to a container along with ability to change its working dir. For example:

- project_root
  |- dip.yml (1)
  |- docker-compose.yml (2)
  |- sub-project-dir
     |- your current directory is here <<<

# dip.yml (1)
environment:
  WORK_DIR: /app/${DIP_WORK_DIR_REL_PATH}

# docker-compose.yml (2)
services:
  app:
    working_dir: ${WORK_DIR:-/app}

cd sub-project-dir
dip run bash -c pwd

returned is /app/sub-project-dir.

Exposes the current user ID (UID). It is useful when you need to run a container with the same user as the host machine. For example:

# dip.yml (1)
environment:
  UID: ${DIP_CURRENT_USER}

# docker-compose.yml (2)
services:
  app:
    image: ruby
    user: ${UID:-1000}

The container will run using the same user ID as your host machine.

Modules are defined as array in modules section of dip.yml, modules are stored in .dip subdirectory of dip.yml directory.

The main purpose of modules is to improve maintainability for a group of projects. Imagine having multiple gems which are managed with dip, each of them has the same commands, so to change one command in dip you need to update all gems individualy.

With modules you can define a group of modules for dip.

For example having setup as this:

# ./dip.yml
modules:
 - sasts
 - rails

...

# ./.dip/sasts.yml
interaction:
  brakeman:
    description: Check brakeman sast
    command: docker run ...

# ./.dip/rails.yml
interaction:
  annotate:
    description: Run annotate command
    service: backend
    command: bundle exec annotate

Will be expanded to:

# resultant configuration
interaction:
  brakeman:
    description: Check brakeman sast
    command: docker run ...
  annotate:
    description: Run annotate command
    service: backend
    command: bundle exec annotate

Imagine .dip to be a submodule so it can be managed only in one place.

If you want to override module command, you can redefine it in dip.yml

# ./dip.yml
modules:
 - sasts

interaction:
  brakeman:
    description: Check brakeman sast
    command: docker run another-image ...

# ./.dip/sasts.yml
interaction:
  brakeman:
    description: Check brakeman sast
    command: docker run some-image ...

Will be expanded to:

# resultant configuration
interaction:
  brakeman:
    description: Check brakeman sast
    command: docker run another-image ...

Nested modules are not supported.

Run commands defined within the interaction section of dip.yml

A command will be executed by specified runner. Dip has three types of them:

• docker compose runner — used when the service option is defined.
• kubectl runner — used when the pod option is defined.
• local runner — used when the previous ones are not defined.

dip run rails c
dip run rake db:migrate

Also, run argument can be omitted

dip rake db:migrate

You can pass in a custom environment variable into a container:

dip VERSION=12352452 rake db:rollback

Use options -p, --publish=[] if you need to additionally publish a container's port(s) to the host unless this behaviour is not configured at dip.yml:

dip run -p 3000:3000 bundle exec rackup config.ru

You can also override docker compose command by passing DIP_COMPOSE_COMMAND if you wish. For example if you want to use mutagen-compose run DIP_COMPOSE_COMMAND=mutagen-compose dip run.

If you want to persist that change you can specify command in compose section of dip.yml :

compose:
  command: mutagen-compose

List all available run commands.

dip ls

bash     # Open the Bash shell in app's container
rails    # Run Rails command
rails s  # Run Rails server at http://localhost:3000

Run commands each by each from provision section of dip.yml

Run Docker Compose commands that are configured according to the application's dip.yml:

dip compose COMMAND [OPTIONS]

dip compose up -d redis

Runs shared Docker Compose services that are used by the current application. Useful for microservices.

There are several official infrastructure services available:

• dip-postgres
• dip-kafka
• dip-nginx

# dip.yml
infra:
  foo:
    git: https://github.com/owner/foo.git
    ref: latest # default, optional
  bar:
    path: ~/path/to/bar

Repositories will be pulled to a ~/.dip/infra folder. For example, for the foo service it would be like this: ~/.dip/infra/foo/latest and clonned with the following command: git clone -b <ref> --single-branch <git> --depth 1.

Available CLI commands:

• dip infra update pulls updates from sources
• dip infra up starts all infra services
• dip infra up -n kafka starts a specific infra service
• dip infra down stops all infra services
• dip infra down -n kafka stops a specific infra service

Run kubectl commands that are configured according to the application's dip.yml:

dip ktl COMMAND [OPTIONS]

STAGE=some dip ktl get pods

Runs ssh-agent container based on https://github.com/whilp/ssh-agent with your ~/.ssh/id_rsa. It creates a named volume ssh_data with ssh socket. An application's docker-compose.yml should contains environment variable SSH_AUTH_SOCK=/ssh/auth/sock and connects to external volume ssh_data.

dip ssh up

docker-compose.yml

services:
  web:
    environment:
      - SSH_AUTH_SOCK=/ssh/auth/sock
    volumes:
      - ssh-data:/ssh:ro

volumes:
  ssh-data:
    external:
      name: ssh_data

if you want to use non-root user you can specify UID like so:

dip ssh up -u 1000

This especially helpful if you have something like this in your docker-compose.yml:

services:
  web:
    user: "1000:1000"

https://github.com/bibendi/dip/releases