Save RUN --mount=type=cache
caches on GitHub Actions or other CI platforms
The BuildKit Cache Dance allows saving RUN --mount=type=cache
caches on GitHub Actions or other CI platforms by extracting the cache from the previous build and injecting it into the current build.
Use cases:
- apt-get (
/var/cache/apt
,/var/lib/apt
) - Go (
/root/.cache/go-build
) - etc.
This reproducible-containers/buildkit-cache-dance
action was forked from
overmindtech/buildkit-cache-dance
(archived on September 2023).
This action be used for "non-reproducible" containers too.
Dockerfile:
FROM ubuntu:22.04
ENV DEBIAN_FRONTEND=noninteractive
RUN \
--mount=type=cache,target=/var/cache/apt,sharing=locked \
--mount=type=cache,target=/var/lib/apt,sharing=locked \
rm -f /etc/apt/apt.conf.d/docker-clean && \
echo 'Binary::apt::APT::Keep-Downloaded-Packages "true";' >/etc/apt/apt.conf.d/keep-cache && \
apt-get update && \
apt-get install -y gcc
Action:
---
name: Build
on:
push:
jobs:
Build:
runs-on: ubuntu-22.04
steps:
- uses: actions/checkout@v4
- uses: docker/setup-buildx-action@v3
id: setup-buildx
- uses: docker/metadata-action@v5
id: meta
with:
images: Build
- name: Cache
uses: actions/cache@v4
id: cache
with:
path: cache-mount
key: cache-mount-${{ hashFiles('Dockerfile') }}
- name: Restore Docker cache mounts
uses: reproducible-containers/buildkit-cache-dance@v3
with:
builder: ${{ steps.setup-buildx.outputs.name }}
cache-dir: cache-mount
dockerfile: Dockerfile
skip-extraction: ${{ steps.cache.outputs.cache-hit }}
- name: Build and push
uses: docker/build-push-action@v5
with:
context: .
cache-from: type=gha
cache-to: type=gha,mode=max
file: Dockerfile
push: ${{ github.event_name != 'pull_request' }}
tags: ${{ steps.meta.outputs.tags }}
labels: ${{ steps.meta.outputs.labels }}
Real-world examples:
- https://github.com/rootless-containers/slirp4netns/blob/v1.2.2/.github/workflows/release.yaml#L18-L36
- https://github.com/containers/fuse-overlayfs/blob/40e0f3c/.github/workflows/release.yaml#L17-L36
If you require more fine grained control you can manually specify a JSON formatted cache-map
. The keys specify the paths on the Docker builder host to use as the bind source and the string value provides the cache mount target
within the Docker build:
- name: Restore Docker cache mounts
uses: reproducible-containers/buildkit-cache-dance@v3
with:
builder: ${{ steps.setup-buildx.outputs.name }}
cache-map: |
{
"var-cache-apt": "/var/cache/apt",
"var-lib-apt": "/var/lib/apt"
}
skip-extraction: ${{ steps.cache.outputs.cache-hit }}
Alternatively, you can provide a JSON object with additional options that should be passed to --mount=type=cache
in the values cache-map
JSON. The target
path must be present in the object as a property.
- name: Restore Docker cache mounts
uses: reproducible-containers/buildkit-cache-dance@v3
with:
builder: ${{ steps.setup-buildx.outputs.name }}
cache-map: |
{
"var-cache-apt": {
"target": "/var/cache/apt",
"id": "1"
},
"var-lib-apt": "/var/lib/apt"
}
skip-extraction: ${{ steps.cache.outputs.cache-hit }}
In other CI systems, you can run the script directly via node
:
curl -LJO https://github.com/reproducible-containers/buildkit-cache-dance/archive/refs/tags/v3.1.0.tar.gz
tar xvf buildkit-cache-dance-3.1.0.tar.gz
During injection:
node ./buildkit-cache-dance-3.1.0/dist/index.js --cache-map '{"var-cache-apt": "/var/cache/apt", "var-lib-apt": "/var/lib/apt"}'
After build during extraction:
node ./buildkit-cache-dance-3.1.0/dist/index.js --extract --cache-map '{"var-cache-apt": "/var/cache/apt", "var-lib-apt": "/var/lib/apt"}'
Here are the available options:
build-cache-dance [options]
Save 'RUN --mount=type=cache' caches on GitHub Actions or other CI platforms
Options:
--extract Extract the cache from the docker container (extract step). Otherwise, inject the cache (main step)
--cache-map The map of actions source to container destination paths for the cache paths
--dockerfile The Dockerfile to use for the auto-discovery of cache-map. Default: 'Dockerfile'
--cache-dir The root directory where cache content is injected from/extracted to when using auto-discovery of the cache-map.
--scratch-dir Where the action is stores some temporary files for its processing. Default: 'scratch'
--skip-extraction Skip the extraction of the cache from the docker container
--builder The name of the buildx builder. Default: 'default'
--help Show this help
v1 follows the original design of overmindtech/buildkit-cache-dance
.
v1 is composed of two actions:
reproducible-containers/buildkit-cache-dance/inject@v1.0.1
reproducible-containers/buildkit-cache-dance/extract@v1.0.1
See the releases/v1
branch.
v2 is composed of the single reproducible-containers/buildkit-cache-dance
action.
Rewrote the action in TypeScript and adds support for cache-map
that gets a string of files that need to be injected as a JSON string. This makes it possible to inject multiple directories in one call and simplifies the usage.
This release also makes it possible to run the script outside GitHub Actions in other CI platforms or locally using command line arguments.
- Thanks to Alexander Pravdin for the basic idea in this comment.
- Thanks to the authors of the original
overmindtech/buildkit-cache-dance
.