feat: add Containerfile for building vllm CPU images #33
Workflow file for this run
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| name: Build, test, and publish vLLM CPU Containers | |
| on: | |
| pull_request: | |
| branches: | |
| - main | |
| - rhoai-v* | |
| - konflux-poc* | |
| types: | |
| - opened | |
| - synchronize | |
| paths: | |
| - 'vllm/Containerfile' | |
| - '.github/actions/**' | |
| - '.github/actions/workflows/vllm-cpu-container.yml' | |
| push: | |
| branches: | |
| - main | |
| - rhoai-v* | |
| paths: | |
| - 'vllm/Containerfile' | |
| - '.github/actions/**' | |
| - '.github/actions/workflows/vllm-cpu-container.yml' | |
| workflow_dispatch: | |
| inputs: | |
| inference_model: | |
| description: 'Inference model to preload onto vLLM image - default is Qwen/Qwen3-0.6B' | |
| type: string | |
| embedding_model: | |
| description: 'Embedding model to preload onto vLLM image - default is ibm-granite/granite-embedding-125m-english' | |
| type: string | |
| env: | |
| REGISTRY: quay.io | |
| IMAGE_NAME: quay.io/opendatahub/vllm-cpu # tags for the image will be added dynamically | |
| jobs: | |
| build-test-push: | |
| runs-on: ubuntu-latest | |
| env: | |
| INFERENCE_MODEL: ${{ github.event.inputs.inference_model || 'Qwen/Qwen3-0.6B' }} | |
| EMBEDDING_MODEL: ${{ github.event.inputs.embedding_model || 'ibm-granite/granite-embedding-125m-english' }} | |
| strategy: | |
| matrix: | |
| platform: [linux/amd64] # TODO: enable other arch once all pip packages are available. | |
| permissions: | |
| contents: read | |
| steps: | |
| - name: Checkout repository | |
| uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # v6.0.1 | |
| - name: Set image tag components | |
| run: | | |
| INFERENCE_TEMP="${INFERENCE_MODEL#*/}" | |
| EMBEDDING_TEMP="${EMBEDDING_MODEL#*/}" | |
| echo "INFERENCE_TAG=${INFERENCE_TEMP%-*}" >> "$GITHUB_ENV" | |
| echo "EMBEDDING_TAG=${EMBEDDING_TEMP%-*}" >> "$GITHUB_ENV" | |
| - name: Install uv | |
| uses: astral-sh/setup-uv@681c641aba71e4a1c380be3ab5e12ad51f415867 # v7.1.6 | |
| with: | |
| python-version: 3.12 | |
| - name: Set up QEMU | |
| uses: docker/setup-qemu-action@c7c53464625b32c7a7e944ae62b3e17d2b600130 # v3.7.0 | |
| - name: Set up Docker Buildx | |
| uses: docker/setup-buildx-action@8d2750c68a42422c14e847fe6c8ac0403b4cbd6f # v3.12.0 | |
| - name: Free disk space | |
| uses: ./.github/actions/free-disk-space | |
| - name: Build image | |
| id: build | |
| uses: docker/build-push-action@263435318d21b8e681c14492fe198d362a7d2c83 # v6.18.0 | |
| with: | |
| context: . | |
| file: vllm/Containerfile | |
| platforms: ${{ matrix.platform }} | |
| push: false | |
| tags: ${{ env.IMAGE_NAME }}:${{ env.INFERENCE_TAG }}-${{ env.EMBEDDING_TAG }} | |
| load: true # needed to load for smoke test | |
| build-args: | | |
| INFERENCE_MODEL=${{ env.INFERENCE_MODEL }} | |
| EMBEDDING_MODEL=${{ env.EMBEDDING_MODEL }} | |
| - name: Setup vllm for inference test | |
| if: github.event_name != 'workflow_dispatch' | |
| id: vllm-inference | |
| uses: ./.github/actions/setup-vllm | |
| env: | |
| VLLM_IMAGE: ${{ env.IMAGE_NAME }}:${{ env.INFERENCE_TAG }}-${{ env.EMBEDDING_TAG }} | |
| VLLM_MODE: 'inference' | |
| - name: Setup vllm for embedding test | |
| if: github.event_name != 'workflow_dispatch' | |
| id: vllm-embedding | |
| uses: ./.github/actions/setup-vllm | |
| env: | |
| VLLM_IMAGE: ${{ env.IMAGE_NAME }}:${{ env.INFERENCE_TAG }}-${{ env.EMBEDDING_TAG }} | |
| VLLM_MODE: 'embedding' | |
| - name: Gather logs and debugging information | |
| if: always() | |
| shell: bash | |
| run: | | |
| # Create logs directory | |
| mkdir -p logs | |
| docker logs vllm-inference > logs/vllm-inference.log 2>&1 || echo "Failed to get vllm-inference logs" > logs/vllm-inference.log | |
| docker logs vllm-embedding > logs/vllm-embedding.log 2>&1 || echo "Failed to get vllm-embedding logs" > logs/vllm-embedding.log | |
| # Gather system information | |
| echo "=== System information ===" | |
| { | |
| echo "Disk usage:" | |
| df -h | |
| echo "Memory usage:" | |
| free -h | |
| echo "Docker images:" | |
| docker images | |
| echo "Docker containers:" | |
| docker ps -a | |
| } > logs/system-info.log 2>&1 | |
| - name: Upload logs as artifacts | |
| if: always() | |
| uses: actions/upload-artifact@b7c566a772e6b6bfb58ed0dc250532a479d7789f # v6.0.0 | |
| with: | |
| name: ci-logs-${{ github.sha }} | |
| path: logs/ | |
| retention-days: 7 | |
| - name: Cleanup vllm containers | |
| if: always() | |
| shell: bash | |
| run: | | |
| docker rm -f vllm-inference vllm-embedding >/dev/null 2>&1 || true | |
| - name: Log in to Quay.io | |
| id: login | |
| if: contains(fromJSON('["push", "workflow_dispatch"]'), github.event_name) | |
| uses: docker/login-action@5e57cd118135c172c3672efd75eb46360885c0ef # v3.6.0 | |
| with: | |
| registry: ${{ env.REGISTRY }} | |
| username: ${{ secrets.QUAY_USERNAME }} | |
| password: ${{ secrets.QUAY_PASSWORD }} | |
| - name: Publish image to Quay.io | |
| id: publish | |
| if: contains(fromJSON('["push", "workflow_dispatch"]'), github.event_name) | |
| uses: docker/build-push-action@263435318d21b8e681c14492fe198d362a7d2c83 # v6.18.0 | |
| with: | |
| context: . | |
| file: vllm/Containerfile | |
| platforms: ${{ matrix.platform }} | |
| push: true | |
| tags: ${{ env.IMAGE_NAME }}:${{ env.INFERENCE_TAG }}-${{ env.EMBEDDING_TAG }} | |
| build-args: | | |
| INFERENCE_MODEL=${{ env.INFERENCE_MODEL }} | |
| EMBEDDING_MODEL=${{ env.EMBEDDING_MODEL }} |