# Useful Docker images
Docker images are like any other software. You should do your own research before using them and this list does not make an exception.
flemay/musketeers has useful tools for a 3 Musketeers project including Docker, Compose, Make, and more. The
flemay/cookiecutterimage in the examples uses this image to build and test the Docker image in GitLab Pipeline. See pattern Docker-in-Docker (DinD)
jwilder/dockerize: There is often a need to wait for a service to start before interacting with it. For instance, waiting for a database container to be ready before running a migration. The image
jwilder/dockerizecan be used to help with this scenario.
dbStart: docker-compose up -d db docker-compose run --rm dockerize -wait tcp://db:3306 -timeout 60s
dockerlint validates your Dockerfiles
shellcheck lints your shell scripts
# Accessing host's localhost from a container
On Windows/Mac, accessing the host localhost is to use the url like
host.docker.internal. This is handy because if you have an application running on
localhost:3000 locally (through container or not), then you can access it
$ curl host.docker.internal:3000.
# Image without Make
One of the patterns is to call Make from Compose. If you want to follow this pattern and your image does not have
make, here are some solutions to address that.
# Use a different image
Often image publishers offer different versions of the application/product. For instance golang has an image based on
alpine which does not have
make. It also has an image based on
stretch which does.
$ docker run --rm golang:alpine make # "exec: \"make\": executable file not found $ docker run --rm golang:stretch make # make: *** No targets specified and no makefile found
# Use Musketeers Docker image
If you only want to call
make with common shell commands, or want to use
zip, then the lightweight Musketeers Docker image is for you.
# Install Make on the fly
Whenever a command runs another command it installs
make and then execute
$ make _target. Depending on how many times a command is run, this may be inefficient as it needs to download
make every time.
$ docker run --rm golang:alpine apk add --update make && make _target
# Build your own image
You may want to build and maintain your own image based on the the image you wanted to use.
FROM node:alpine RUN apk add --update make ...
# Docker development is slow
Mounting volumes with Docker on Mac or Windows can be slow. For instance, developing a rails application. A handy tool which can help solve this problem is docker-sync
On Mac, using the
native_osx strategy can also help. The Docker Compose file would look like the following:
yourservice: image: animage volumes: - app-sync:/opt/app:nocopy ... volumes: # this volume is created by docker-sync. See docker-sync.yml for the config app-sync: external: true
This would work well on Windows/Mac but what about Linux? Either docker-sync is still used, which uses the native strategy and would not sync, or you use an environment variable which sets the volume: