DockerMA facilitates building multi-arch containers with minimal fuss
DockerMA facilitates building multi-arch containers with minimal fuss.
What does it do?
dockerma aims to be a drop-in replacement for docker build and docker push to handle building what is often referred to as “multi-arch” images. A “multi-arch” image is actually a manifest, listing which actual image to use for a given architecture. The way to create these are still somewhat cumbersome, and considered experimental, especially if you are using a base image that is already “multi-arch”.
dockerma build works by reading your Dockerfile and looking for a specially formatted comment that selects which architectures you want to build. It will then create a manipulated version of your Dockerfile, to inject some tools for cross-building, and then build architecture specific images. If you have selected amd64 and arm, and build myapp:v1.0, the build will produce myapp:v1.0-amd64 and myapp:v1.0-arm.
dockerma push will take the architecture specific images that was built previously, and push them to the remote registry. It will then create a manifest, listing each of the images for the wanted architectures, and push the manifest under the requested tag. This will create a “multi-arch” image.
Supported architectures is based on what is available in the tooling used. The cross-building tools are from https://github.com/monsonnl/qemu-wrap-build-files, and at this writing only supports amd64, arm and arm64. If you want dockerma to support other architectures, you need to persuade qemu-wrap-build-files to support it first.
How do I use it?
Since dockerma aims to be a drop-in replacement, using it should be fairly simple.
The first step is to add the dockerma comment to your Dockerfile:
# dockerma archs:arm,amd64,arm64:
When you previously used docker build -t myapp:latest -t myapp:v1.0 . to build latest and v1.0 tags for your current architecture, you will now simply replace docker with dockerma:
dockerma build -t myapp:latest -t myapp:v1.0 .
Similary, where you used docker push -t myapp:latest, you now use dockerma:
dockerma push -t myapp:latest
What are the requirements?
- You need to enable experimental features in docker. This can be done by setting DOCKER_CLI_EXPERIMENTAL to enabled, or changing your docker configuration. Consult the docker documentation for details.
- Your build host must support running qemu-<arch>-static.
Are there any downsides to this approach?
- All images will have some additional files from the cross-build tooling, located in /cross-build.
Releasing a new version
- Tag the new version: hg tag v1.2.3
- Push the tag to Bitbucket
- Watch the pipeline publish the new version to PyPI
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size & hash SHA256 hash help||File type||Python version||Upload date|
|dockerma-0.4.1-py2-none-any.whl (9.6 kB) Copy SHA256 hash SHA256||Wheel||py2|
|dockerma-0.4.1-py3-none-any.whl (9.6 kB) Copy SHA256 hash SHA256||Wheel||py3|
|dockerma-0.4.1.tar.gz (12.8 kB) Copy SHA256 hash SHA256||Source||None|