updates from other PR

Author: nv-anants

.github/workflows/build-and-test-vllm.yml

@@ -15,6 +15,11 @@
 
 name: NVIDIA Github Validation - vLLM
 
+# Do not cancel merges into main
+concurrency:
+  group: pr-build-test-vllm-${{ github.event.number || github.run_id }}
+  cancel-in-progress: ${{ github.ref_name != 'main' }}
+
 on:
   push:
     branches:
@@ -23,8 +28,7 @@ on:
 
 jobs:
   build-test:
-    runs-on:
-        group: GPU_runners
+    runs-on: gpu-l40-runners
     name: Build and Test - vLLM
     env:
       CONTAINER_ID: test_${{ github.run_id }}_${{ github.run_attempt }}_${{ github.job }}_vllm
@@ -51,9 +55,9 @@ jobs:
           ./container/build.sh --tag ${{ steps.define_image_tag.outputs.image_tag }} --target runtime --framework vllm
       - name: Run pytest
         env:
-          PYTEST_MARKS: "e2e and vllm and gpu_1 and serve"
+          PYTEST_MARKS: "e2e and vllm and gpu_1 and not slow"
         run: |
-          docker run --gpus all-v ${{ github.workspace }}:/workspace -w /workspace --name ${{ env.CONTAINER_ID }}_pytest ${{ steps.define_image_tag.outputs.image_tag }} bash -c "pytest --basetemp=/tmp --junitxml=${{ env.PYTEST_XML_FILE }} -m \"${{ env.PYTEST_MARKS }}\""
+          docker run --gpus all -v ${{ github.workspace }}:/workspace -w /workspace --name ${{ env.CONTAINER_ID }}_pytest ${{ steps.define_image_tag.outputs.image_tag }} bash -c "pytest --basetemp=/tmp --junitxml=${{ env.PYTEST_XML_FILE }} -m \"${{ env.PYTEST_MARKS }}\""
       - name: Cleanup - Remove dynamo directory
         if: always()
         run: |

tests/serve/test_vllm.py

@@ -302,8 +302,6 @@ def vllm_config_test(request):
 
 
 @pytest.mark.e2e
-@pytest.mark.slow
-@pytest.mark.serve
 def test_serve_deployment(vllm_config_test, request, runtime_services):
     """
     Test dynamo serve deployments with different graph configurations.