diff --git a/.github/workflows/integration-test.yml b/.github/workflows/integration-test.yml index d71afe885a..96dd180108 100644 --- a/.github/workflows/integration-test.yml +++ b/.github/workflows/integration-test.yml @@ -1212,28 +1212,31 @@ jobs: terraform destroy -auto-approve && exit 1 fi - - name: Apply GPU burner YAML - run: kubectl apply -f ./gpuBurner.yaml - - - name: Create NVIDIA device plugin - run: kubectl create -f https://raw.githubusercontent.com/NVIDIA/k8s-device-plugin/v0.15.0/deployments/static/nvidia-device-plugin.yml - - name: Patch the CloudWatch Agent image and restart CloudWatch pods run: | kubectl patch amazoncloudwatchagents -n amazon-cloudwatch cloudwatch-agent --type='json' -p='[{"op": "replace", "path": "/spec/image", "value": ${{ secrets.AWS_ECR_PRIVATE_REGISTRY }}/${{ env.ECR_INTEGRATION_TEST_REPO }}:${{ github.sha }}}]' kubectl delete pods --all -n amazon-cloudwatch sleep 10 kubectl wait --for=condition=Ready pod --all -n amazon-cloudwatch + - name: Get EKS Cluster Name + id: get-eks-cluster-name + run: | + cd terraform/eks/addon/gpu + CLUSTER_NAME=$(terraform output -raw eks_cluster_name) + echo "::set-output name=eks_cluster_name::$CLUSTER_NAME" + - name: Run tests run: | - if go test ${var.test_dir} -eksClusterName ${aws_eks_cluster.this.name} -computeType=EKS -v -eksDeploymentStrategy=DAEMON -eksGpuType=nvidia; then + CLUSTER_NAME=${{ steps.get-eks-cluster-name.outputs.eks_cluster_name }} + if go test test/gpu -eksClusterName $CLUSTER_NAME -computeType=EKS -v -eksDeploymentStrategy=DAEMON -eksGpuType=nvidia; then echo "Tests passed" else echo "Tests failed" exit 1 fi - + + - name: Terraform destroy if: ${{ cancelled() || failure() }} uses: nick-fields/retry@v2