Merge branch 'main' into vllm-0.15.1-update

Author: junpuf

.github/config/vllm-0.10.2-rayserve.yml

@@ -12,7 +12,7 @@ common:
   framework_version: "0.10.2"
   job_type: "general"
   python_version: "py312"
-  cuda_version: "cu129"
+  cuda_version: "cu128"
   os_version: "ubuntu22.04"
   customer_type: "rayserve_ec2"
   arch_type: "x86"
@@ -26,5 +26,4 @@ release:
   public_registry: true
   private_registry: false
   enable_soci: true
-  source_stage: beta # private for gamma test
-  target_stage: release # gamma for gamma test
+  environment: production  # gamma or production

.github/workflows/reusable-release-image.yml

@@ -14,12 +14,8 @@ on:
         description: 'Release specification YAML content'
         required: true
         type: string
-      source-stage:
-        description: 'Source stage (e.g., private, beta)'
-        required: true
-        type: string
-      target-stage:
-        description: 'Target stage (e.g., gamma, release)'
+      environment:
+        description: 'Deployment environment (gamma or production)'
         required: true
         type: string
       aws-region:
@@ -38,7 +34,27 @@ on:
         value: ${{ jobs.step1-publish-images.outputs.run-identifier }}
 
 jobs:
+  validate-release:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Validate release context
+        run: |
+          ENVIRONMENT="${{ inputs.environment }}"
+
+          # Validate environment input
+          if [[ "${ENVIRONMENT}" != "gamma" && "${ENVIRONMENT}" != "production" ]]; then
+            echo "❌ ERROR: Invalid environment '${ENVIRONMENT}'"
+            echo "Valid environments: gamma, production"
+            exit 1
+          fi
+
+          echo "✅ Release context validation passed"
+          echo "Ref: ${{ github.ref }}"
+          echo "Environment: ${ENVIRONMENT}"
+
   step1-publish-images:
+    needs: [validate-release]
+    environment: ${{ inputs.environment }}
     runs-on:
       - codebuild-runner-${{ github.run_id }}-${{ github.run_attempt }}
         fleet:${{ inputs.runner-fleet }}
@@ -107,13 +123,16 @@ jobs:
           echo "RUN_IDENTIFIER=${RUN_IDENTIFIER}" >> ${GITHUB_ENV}
           echo "METADATA_FILE=${METADATA_FILE}" >> ${GITHUB_ENV}
           echo "REGION=${{ inputs.aws-region }}" >> ${GITHUB_ENV}
-          echo "SOURCE_STAGE=${{ inputs.source-stage }}" >> ${GITHUB_ENV}
-          echo "TARGET_STAGE=${{ inputs.target-stage }}" >> ${GITHUB_ENV}
+          echo "SOURCE_STAGE=${{ vars.SOURCE_STAGE }}" >> ${GITHUB_ENV}
+          echo "TARGET_STAGE=${{ vars.TARGET_STAGE }}" >> ${GITHUB_ENV}
 
           echo "run-identifier=${RUN_IDENTIFIER}" >> ${GITHUB_OUTPUT}
           echo "metadata-file-name=release_metadata_${RUN_IDENTIFIER}.dict" >> ${GITHUB_OUTPUT}
 
           echo "Generated run identifier: ${RUN_IDENTIFIER}"
+          echo "Environment: ${{ inputs.environment }}"
+          echo "Source stage: ${{ vars.SOURCE_STAGE }}"
+          echo "Target stage: ${{ vars.TARGET_STAGE }}"
 
       - name: Publish DLC Images
         id: publish
@@ -161,6 +180,7 @@ jobs:
 
   step2-generate-info:
     needs: [step1-publish-images]
+    environment: ${{ inputs.environment }}
     if: needs.step1-publish-images.outputs.images-published == '1'
     runs-on:
       - codebuild-runner-${{ github.run_id }}-${{ github.run_attempt }}
@@ -189,8 +209,8 @@ jobs:
           echo "RUN_IDENTIFIER=${RUN_IDENTIFIER}" >> ${GITHUB_ENV}
           echo "METADATA_FILE=${METADATA_FILE}" >> ${GITHUB_ENV}
           echo "REGION=${{ inputs.aws-region }}" >> ${GITHUB_ENV}
-          echo "SOURCE_STAGE=${{ inputs.source-stage }}" >> ${GITHUB_ENV}
-          echo "TARGET_STAGE=${{ inputs.target-stage }}" >> ${GITHUB_ENV}
+          echo "SOURCE_STAGE=${{ vars.SOURCE_STAGE }}" >> ${GITHUB_ENV}
+          echo "TARGET_STAGE=${{ vars.TARGET_STAGE }}" >> ${GITHUB_ENV}
 
       - name: Generate Release Information
         run: |
@@ -214,6 +234,7 @@ jobs:
 
   step3-publish-notifications:
     needs: [step1-publish-images, step2-generate-info]
+    environment: ${{ inputs.environment }}
     if: needs.step1-publish-images.outputs.images-published == '1'
     runs-on:
       - codebuild-runner-${{ github.run_id }}-${{ github.run_attempt }}
@@ -233,8 +254,8 @@ jobs:
           RUN_IDENTIFIER="${{ needs.step1-publish-images.outputs.run-identifier }}"
           echo "RUN_IDENTIFIER=${RUN_IDENTIFIER}" >> ${GITHUB_ENV}
           echo "REGION=${{ inputs.aws-region }}" >> ${GITHUB_ENV}
-          echo "SOURCE_STAGE=${{ inputs.source-stage }}" >> ${GITHUB_ENV}
-          echo "TARGET_STAGE=${{ inputs.target-stage }}" >> ${GITHUB_ENV}
+          echo "SOURCE_STAGE=${{ vars.SOURCE_STAGE }}" >> ${GITHUB_ENV}
+          echo "TARGET_STAGE=${{ vars.TARGET_STAGE }}" >> ${GITHUB_ENV}
 
       - name: Publish Release Information
         run: |
@@ -248,4 +269,4 @@ jobs:
           echo ""
           echo "=========================================="
           echo "🎉 All release steps completed successfully!"
-          echo "=========================================="
+          echo "=========================================="

.github/workflows/vllm-rayserver-auto-release.yml

@@ -1,7 +1,8 @@
 name: vLLM_RAYSERVE_AUTO_RELEASE
 
 on:
-
+  # schedule run must be on the default branch
+  # https://docs.github.com/en/actions/reference/workflows-and-actions/events-that-trigger-workflows#schedule
   schedule:
     # Runs at 10:00 AM PST/PDT on Monday and Wednesday
     # Note: GitHub Actions uses UTC time. PST is UTC-8, PDT is UTC-7
@@ -318,8 +319,7 @@ jobs:
     with:
       source-image-uri: ${{ needs.build-vllm-rayserve-image.outputs.ci-image }}
       release-spec: ${{ needs.generate-rayserve-release-spec.outputs.release-spec }}
-      source-stage: ${{ fromJson(needs.load-config.outputs.rayserve-config).release.source_stage }}
-      target-stage: ${{ fromJson(needs.load-config.outputs.rayserve-config).release.target_stage }}
+      environment: ${{ fromJson(needs.load-config.outputs.rayserve-config).release.environment }}
       aws-region: ${{ vars.AWS_REGION }}
       runner-fleet: default-runner
 

docs/DEVELOPMENT.md

@@ -239,6 +239,46 @@ cd docs && mkdocs serve
 
 ______________________________________________________________________
 
+## Local Documentation Development
+
+### Generation Only (No Server)
+
+Run `main.py` to generate documentation without serving:
+
+```bash
+cd docs/src && python main.py --verbose
+```
+
+This automatically clones `tutorials/` repository and generates markdown files in `reference/` and `releasenotes/` directories without starting a web server.
+
+### Serving Locally
+
+Use `mkdocs serve` to automatically clone `tutorials/` and generate documentation in `reference/` and `releasenotes/` and serve the website:
+
+```bash
+cd docs && mkdocs serve
+```
+
+The site is typically available at `http://127.0.0.1:8000/deep-learning-containers/` - check the command output for the actual URL.
+
+### Live Reload
+
+Enable automatic reload on content changes:
+
+```bash
+mkdocs serve --livereload
+```
+
+**Note:** Live reload only detects changes to:
+
+- Markdown file content
+- `.nav.yml` content
+- `mkdocs.yml` content
+
+Live reload does **not** detect changes requiring documentation regeneration (e.g., image config YAML files, templates). To regenerate documentation, stop the server (`Ctrl+C`) and rerun `mkdocs serve`.
+
+______________________________________________________________________
+
 ## Troubleshooting
 
 | Error | Solution |

docs/src/data/template/image-template.yml

@@ -6,7 +6,7 @@
 # =============================================================================
 
 # -----------------------------------------------------------------------------
-# Required Fields
+# Common Fields
 # -----------------------------------------------------------------------------
 
 # Framework display name (e.g., "PyTorch", "TensorFlow")
@@ -18,9 +18,6 @@ version: "2.9"
 # Accelerator type: gpu, cpu, or neuronx
 accelerator: gpu
 
-# Python version identifier (e.g., py310, py311, py312)
-python: py312
-
 # Target platform: ec2 or sagemaker
 platform: ec2
 
@@ -37,6 +34,9 @@ tags:
 # For example, if you require a column called "foo",
 # add foo field here and in your desired table YAML file.
 
+# Python version identifier (e.g., py310, py311, py312)
+python: py312
+
 # CUDA version identifier (for GPU images)
 cuda: cu130
 

docs/src/utils.py

@@ -85,6 +85,11 @@ def parse_version(version_str: str | None) -> Version:
 def clone_git_repository(git_repository: str, target_dir: str | Path) -> None:
     """Clone a git repository to target directory if it doesn't exist."""
     if os.path.exists(target_dir):
+        LOGGER.info(
+            f"{git_repository} already exists within {target_dir}"
+            "If you want to pull the latest commit of the repository, either delete the repository and re-run generation"
+            f"or manually run `git pull` within the {target_dir}"
+        )
         return
     subprocess.run(["git", "clone", "--depth", "1", git_repository, target_dir], check=True)