[RFC] Bakefile Multi Arch builds

.github/workflows/build.yml

@@ -16,109 +16,49 @@ jobs:
       release: "2024.9.1"
       defaultpython: "3.10"
 
-    strategy:
-      fail-fast: false
-      matrix:
-        python: ["3.10", "3.11", "3.12"]
-        image:
-          - tag: "daskdev/dask"
-            context: "./base"
-          - tag: "ghcr.io/dask/dask"
-            context: "./base"
-          - tag: "daskdev/dask-notebook"
-            context: "./notebook"
-          - tag: "ghcr.io/dask/dask-notebook"
-            context: "./notebook"
-
     steps:
       - name: Checkout
         uses: actions/checkout@v2
 
       - name: Set up QEMU
         uses: docker/setup-qemu-action@v1
-
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v1
-        with:
-          driver: docker
-
-      - name: Install Container Canary
-        run: |
-          LATEST_RELEASE=$(curl --retry 6 --retry-delay 10 -s https://api.github.com/repos/NVIDIA/container-canary/releases/latest | jq -r ".assets[] | select(.name | test(\"canary_linux_amd64$\")) | .browser_download_url")
-          curl --retry 6 --retry-delay 10 -sSL $LATEST_RELEASE > /usr/local/bin/canary
-          chmod +x /usr/local/bin/canary
-          canary version
-
-      - name: Login to DockerHub
-        uses: docker/login-action@v1
-        if: github.repository == 'dask/dask-docker' && github.event_name == 'push'
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
+      -
+        name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      #- name: Install Container Canary
+      #  run: |
+      #    LATEST_RELEASE=$(curl --retry 6 --retry-delay 10 -s https://api.github.com/repos/NVIDIA/container-canary/releases/latest | jq -r ".assets[] | select(.name | test(\"canary_linux_amd64$\")) | .browser_download_url")
+      #    curl --retry 6 --retry-delay 10 -sSL $LATEST_RELEASE > /usr/local/bin/canary
+      #    chmod +x /usr/local/bin/canary
+      #    canary version
+
+      #- name: Login to DockerHub
+      #  uses: docker/login-action@v1
+      #  if: github.repository == 'dask/dask-docker' && github.event_name == 'push'
+      #  with:
+      #    username: ${{ secrets.DOCKERHUB_USERNAME }}
+      #    password: ${{ secrets.DOCKERHUB_TOKEN }}
 
       - name: Login to GitHub Container Registry
         uses: docker/login-action@v1
-        if: github.repository == 'dask/dask-docker' && github.event_name == 'push'
+      #  if: github.repository == 'dask/dask-docker' && github.event_name == 'push'
         with:
           registry: ghcr.io
           username: ${{ github.actor }}
           password: ${{ secrets.GITHUB_TOKEN }}
 
-      - name: Generate tags
-        id: tags
-        env:
-          image: ${{ matrix.image.tag }}
-          python: ${{ matrix.python }}
-          stable: ${{ startsWith(github.ref, 'refs/tags') }}
-        run: |
-          if [ "$stable" == "true" ]; then
-            tag="${image}:${release}-py${python}"
-            tags=${image}:latest-py${python},$tag
-            if [ "$python" == "$defaultpython" ]; then
-              tags=${image}:latest,${image}:${release},$tags
-            fi
-          else
-            tag="${image}:dev-py${python}"
-            tags=$tag
-            if [ "$python" == "$defaultpython" ]; then
-              tags=${image}:dev,$tags
-            fi
-          fi
-
-          echo "tags=${tags}" >> $GITHUB_OUTPUT
-          echo "tag=${tag}" >> $GITHUB_OUTPUT
-
-      - name: Checkout upstream Jupyter Lab image repo
-        uses: actions/checkout@v2
-        if: contains(matrix.image.tag, 'dask-notebook')
-        with:
-          repository: jupyter/docker-stacks
-          ref: main
-          path: docker-stacks
-
-      - name: Build upstream Jupyter Lab image
-        uses: docker/build-push-action@v2
-        if: contains(matrix.image.tag, 'dask-notebook')
-        with:
-          context: ./docker-stacks/images/docker-stacks-foundation
-          push: false
-          load: true
-          platforms: linux/amd64
-          tags: daskdev/docker-stacks-foundation:lab
-          build-args: |
-            PYTHON_VERSION=${{ matrix.python }}
-
       - name: Build and push
-        uses: docker/build-push-action@v2
+        uses: docker/bake-action@v4
         with:
-          context: ${{ matrix.image.context }}
-          push: ${{ github.repository == 'dask/dask-docker' && github.event_name == 'push' }}
-          platforms: linux/amd64
-          tags: ${{ steps.tags.outputs.tags }}
-          build-args: |
-            python=${{ matrix.python }}
-            release=${{ env.release }}
-            base=daskdev
+          push: true
+          targets: dask
+          files: bakefile.hcl
+        env:
+          STABLE: ${{ startsWith(github.ref, 'refs/tags') }}
+          RELEASE: ${{ env.release }}
+          DEFAULT_PY_VERSION: ${{ env.defaultpython }}
+
 
       - name: Validate Jupyter Lab image for Binder
         if: contains(matrix.image.tag, 'dask-notebook')
@@ -128,5 +68,5 @@ jobs:
         if: contains(matrix.image.tag, 'dask-notebook')
         run: canary validate --file https://github.com/NVIDIA/container-canary/raw/main/examples/kubeflow.yaml ${{ steps.tags.outputs.tag }}
 
-      - name: Report
-        run: echo Built ${{ steps.tags.outputs.tags }}
+      #- name: Report
+      #  run: echo Built ${{ steps.tags.outputs.tags }}

Makefile

@@ -0,0 +1,8 @@
+
+build-docker:
+	docker run --rm --privileged multiarch/qemu-user-static --reset -p yes
+	docker buildx create --name dask-builder || true
+	docker buildx bake --file bakefile.hcl --builder=dask-builder dask
+
+print-docker:
+	docker buildx bake --file bakefile.hcl dask --print

bakefile.hcl

@@ -0,0 +1,142 @@
+variable "GH_REGISTRY" {
+    #default = "ghcr.io/dask"
+    default = "ghcr.io/ryanolee"
+}
+
+variable "DOCKER_HUB_REGISTRY" {
+    #default = "daskdev"
+}
+
+variable "RELEASE" {
+    default = "2024.9.1"
+}
+
+variable "STABLE" {
+    default = "true"
+}
+
+variable "DEFAULT_PY_VERSION" {
+    default = "3.10"
+}
+
+function "tags" {
+    params = [
+        image,
+        label
+    ]
+    result = [
+        "${GH_REGISTRY}/${image}:${label}",
+        #"${DOCKER_HUB_REGISTRY}/${image}:${label}"
+    ]
+}
+
+function "tag_stable" {
+    params = [
+        image,
+        release,
+        py_version
+    ]
+    result = concat(
+        tags(image, "${release}-py${py_version}"),
+        tags(image, "latest-py${py_version}"),
+        equal(py_version, DEFAULT_PY_VERSION) ? tags(image, "latest") : [],
+        equal(py_version, DEFAULT_PY_VERSION) ? tags(image, release) : []
+    )
+}
+
+function "tag_dev" {
+    params = [
+        image,
+        py_version
+    ]
+    result = concat(
+        tags(image, "dev-py${py_version}"),
+        equal(py_version, DEFAULT_PY_VERSION) ? tags(image, "dev") : []
+    )
+}
+
+function "tag" {
+    params = [
+        image,
+        release,
+        py_version
+    ]
+    result = equal(STABLE, "true") ? tag_stable(image, release, py_version) : tag_dev(image, py_version)
+}
+
+group "dask" {
+    targets = ["base", "notebook"]
+}
+
+target "base" {
+    context = "./base"
+    dockerfile = "Dockerfile"
+    platforms = ["linux/amd64", "linux/arm64"]
+    matrix = {
+        py_version = ["3.10", "3.11", "3.12"]
+    }
+    args = {
+        python = py_version
+        release = RELEASE
+    }
+
+    name = "dask-${replace(py_version, ".", "")}"
+    tags = tag(
+        "dask",
+        RELEASE,
+        py_version
+    )
+}
+
+target "notebook-base" {
+    context = "https://github.com/jupyter/docker-stacks.git#main:images/docker-stacks-foundation"
+    dockerfile = "Dockerfile"
+    platforms = ["linux/amd64", "linux/arm64"]
+    matrix = {
+        py_version = ["3.12"]
+    }
+    args = {
+        PYTHON_VERSION = py_version
+    }
+    name = "notebook-base-${replace(py_version, ".", "")}"
+}
+
+# Pre 311 builds do not work due to https://github.com/jupyter/docker-stacks/issues/2146#issuecomment-2382315848
+# patch works for now. This can be collapsed into "notebook-base" once a fix has been released
+target "notebook-base-pre-311" {
+    context = "https://github.com/jupyter/docker-stacks.git#5365b9f79fa4ffbb20f10133cc6ac5bec5046302:images/docker-stacks-foundation"
+    dockerfile = "Dockerfile"
+    platforms = ["linux/amd64", "linux/arm64"]
+    matrix = {
+        py_version = ["3.10", "3.11"]
+    }
+    args = {
+        PYTHON_VERSION = py_version
+    }
+    name = "notebook-base-${replace(py_version, ".", "")}"
+}
+
+target "notebook" {
+    context = "./notebook"
+    dockerfile = "Dockerfile"
+    platforms = ["linux/amd64", "linux/arm64"]
+    matrix = {
+        py_version = ["3.12", "3.11", "3.10"]
+    }
+    args = {
+        python = py_version
+        release = RELEASE
+        base = "baseapp"
+    }
+
+    contexts = {
+        baseapp = "target:notebook-base-${replace(py_version, ".", "")}"
+    }
+
+    name = "dask-notebook-${replace(py_version, ".", "")}"
+    tags = tag(
+        "dask-notebook",
+        RELEASE,
+        py_version
+    )
+}

base/Dockerfile

@@ -10,14 +10,15 @@ ENV PATH /opt/conda/bin:$PATH
 ENV PYTHON_VERSION=${python}
 ENV DASK_VERSION=${release}
 
-RUN mamba install -y \
-    "mamba>=0.27.0" \
-    python=${PYTHON_VERSION} \
-    nomkl \
-    cmake \
-    dask=${DASK_VERSION} \
-    cachey \
-    streamz \
+RUN export G_SLICE=always-malloc \
+    && mamba install -y \
+        "mamba>=0.27.0" \
+        python=${PYTHON_VERSION} \
+        nomkl \
+        cmake \
+        dask=${DASK_VERSION} \
+        cachey \
+        streamz \
     && mamba clean -tipy \
     && find /opt/conda/ -type f,l -name '*.a' -delete \
     && find /opt/conda/ -type f,l -name '*.pyc' -delete \

notebook/Dockerfile

@@ -1,6 +1,6 @@
 ARG base
 
-FROM $base/docker-stacks-foundation:lab
+FROM $base
 
 ARG release
 ARG python
@@ -27,19 +27,20 @@ RUN MAGICARCH=$(dpkg --print-architecture) && \
 
 USER $NB_USER
 
-RUN mamba install -y \
-    python=${PYTHON_VERSION} \
-    nomkl \
-    cytoolz \
-    dask==${DASK_VERSION} \
-    lz4 \
-    numpy \
-    pandas \
-    ipywidgets \
-    cachey \
-    streamz \
-    jupyterlab \
-    "dask-labextension>=5" \
+RUN export G_SLICE=always-malloc \
+    && mamba install -y \
+        python=${PYTHON_VERSION} \
+        nomkl \
+        cytoolz \
+        dask==${DASK_VERSION} \
+        lz4 \
+        numpy \
+        pandas \
+        ipywidgets \
+        cachey \
+        streamz \
+        jupyterlab \
+        "dask-labextension>=5" \
     && mamba clean -tipy \
     && jupyter lab clean \
     && find /opt/conda/ -type f,l -name '*.a' -delete \
@@ -81,4 +82,3 @@ HEALTHCHECK  --interval=15s --timeout=3s --start-period=5s --retries=3 \
 USER $NB_USER
 
 ENTRYPOINT ["tini", "--", "/usr/bin/prepare.sh"]
-CMD ["jupyter", "lab", '--NotebookApp.base_url="${NB_PREFIX:-/}"', '${JUPYTERLAB_ARGS}']

notebook/prepare.sh

@@ -35,4 +35,4 @@ fi
 
 
 # Execute the jupyterlab as specified.
-exec start.sh $@
+exec start.sh jupyter lab --NotebookApp.base_url="${NB_PREFIX:-/}" ${JUPYTERLAB_ARGS}