|
| 1 | +# rayjob e2e tests workflow for CodeFlare-SDK |
| 2 | +name: rayjob-e2e |
| 3 | + |
| 4 | +on: |
| 5 | + pull_request: |
| 6 | + branches: |
| 7 | + - main |
| 8 | + - "release-*" |
| 9 | + - ray-jobs-feature |
| 10 | + paths-ignore: |
| 11 | + - "docs/**" |
| 12 | + - "**.adoc" |
| 13 | + - "**.md" |
| 14 | + - "LICENSE" |
| 15 | + |
| 16 | +concurrency: |
| 17 | + group: ${{ github.head_ref }}-${{ github.workflow }} |
| 18 | + cancel-in-progress: true |
| 19 | + |
| 20 | +env: |
| 21 | + CODEFLARE_OPERATOR_IMG: "quay.io/project-codeflare/codeflare-operator:dev" |
| 22 | + |
| 23 | +jobs: |
| 24 | + kubernetes-rayjob: |
| 25 | + runs-on: gpu-t4-4-core |
| 26 | + |
| 27 | + steps: |
| 28 | + - name: Checkout code |
| 29 | + uses: actions/checkout@v4 |
| 30 | + with: |
| 31 | + submodules: recursive |
| 32 | + |
| 33 | + - name: Checkout common repo code |
| 34 | + uses: actions/checkout@v4 |
| 35 | + with: |
| 36 | + repository: "project-codeflare/codeflare-common" |
| 37 | + ref: "main" |
| 38 | + path: "common" |
| 39 | + |
| 40 | + - name: Checkout CodeFlare operator repository |
| 41 | + uses: actions/checkout@v4 |
| 42 | + with: |
| 43 | + repository: project-codeflare/codeflare-operator |
| 44 | + path: codeflare-operator |
| 45 | + |
| 46 | + - name: Set Go |
| 47 | + uses: actions/setup-go@v5 |
| 48 | + with: |
| 49 | + go-version-file: "./codeflare-operator/go.mod" |
| 50 | + cache-dependency-path: "./codeflare-operator/go.sum" |
| 51 | + |
| 52 | + - name: Set up gotestfmt |
| 53 | + uses: gotesttools/gotestfmt-action@v2 |
| 54 | + with: |
| 55 | + token: ${{ secrets.GITHUB_TOKEN }} |
| 56 | + |
| 57 | + - name: Set up specific Python version |
| 58 | + uses: actions/setup-python@v5 |
| 59 | + with: |
| 60 | + python-version: "3.11" |
| 61 | + cache: "pip" # caching pip dependencies |
| 62 | + |
| 63 | + - name: Setup NVidia GPU environment for KinD |
| 64 | + uses: ./common/github-actions/nvidia-gpu-setup |
| 65 | + |
| 66 | + - name: Setup and start KinD cluster |
| 67 | + uses: ./common/github-actions/kind |
| 68 | + with: |
| 69 | + worker-nodes: 1 |
| 70 | + |
| 71 | + - name: Install NVidia GPU operator for KinD |
| 72 | + uses: ./common/github-actions/nvidia-gpu-operator |
| 73 | + |
| 74 | + - name: Deploy CodeFlare stack |
| 75 | + id: deploy |
| 76 | + run: | |
| 77 | + cd codeflare-operator |
| 78 | + echo Setting up CodeFlare stack |
| 79 | + make setup-e2e KUEUE_VERSION=v0.13.4 KUBERAY_VERSION=v1.4.0 |
| 80 | + echo Deploying CodeFlare operator |
| 81 | + make deploy -e IMG="${CODEFLARE_OPERATOR_IMG}" -e ENV="e2e" |
| 82 | + kubectl wait --timeout=120s --for=condition=Available=true deployment -n openshift-operators codeflare-operator-manager |
| 83 | + cd .. |
| 84 | +
|
| 85 | + - name: Add user to KinD |
| 86 | + uses: ./common/github-actions/kind-add-user |
| 87 | + with: |
| 88 | + user-name: sdk-user |
| 89 | + |
| 90 | + - name: Configure RBAC for sdk user with limited permissions |
| 91 | + run: | |
| 92 | + kubectl create clusterrole list-ingresses --verb=get,list --resource=ingresses |
| 93 | + kubectl create clusterrolebinding sdk-user-list-ingresses --clusterrole=list-ingresses --user=sdk-user |
| 94 | + kubectl create clusterrole namespace-creator --verb=get,list,create,delete,patch --resource=namespaces |
| 95 | + kubectl create clusterrolebinding sdk-user-namespace-creator --clusterrole=namespace-creator --user=sdk-user |
| 96 | + kubectl create clusterrole raycluster-creator --verb=get,list,create,delete,patch --resource=rayclusters |
| 97 | + kubectl create clusterrolebinding sdk-user-raycluster-creator --clusterrole=raycluster-creator --user=sdk-user |
| 98 | + kubectl create clusterrole rayjob-creator --verb=get,list,create,delete,patch --resource=rayjobs |
| 99 | + kubectl create clusterrolebinding sdk-user-rayjob-creator --clusterrole=rayjob-creator --user=sdk-user |
| 100 | + kubectl create clusterrole rayjob-status-reader --verb=get,list,patch,update --resource=rayjobs/status |
| 101 | + kubectl create clusterrolebinding sdk-user-rayjob-status-reader --clusterrole=rayjob-status-reader --user=sdk-user |
| 102 | + kubectl create clusterrole appwrapper-creator --verb=get,list,create,delete,patch --resource=appwrappers |
| 103 | + kubectl create clusterrolebinding sdk-user-appwrapper-creator --clusterrole=appwrapper-creator --user=sdk-user |
| 104 | + kubectl create clusterrole resourceflavor-creator --verb=get,list,create,delete --resource=resourceflavors |
| 105 | + kubectl create clusterrolebinding sdk-user-resourceflavor-creator --clusterrole=resourceflavor-creator --user=sdk-user |
| 106 | + kubectl create clusterrole clusterqueue-creator --verb=get,list,create,delete,patch --resource=clusterqueues |
| 107 | + kubectl create clusterrolebinding sdk-user-clusterqueue-creator --clusterrole=clusterqueue-creator --user=sdk-user |
| 108 | + kubectl create clusterrole localqueue-creator --verb=get,list,create,delete,patch --resource=localqueues |
| 109 | + kubectl create clusterrolebinding sdk-user-localqueue-creator --clusterrole=localqueue-creator --user=sdk-user |
| 110 | + kubectl create clusterrole list-secrets --verb=get,list --resource=secrets |
| 111 | + kubectl create clusterrolebinding sdk-user-list-secrets --clusterrole=list-secrets --user=sdk-user |
| 112 | + kubectl create clusterrole pod-creator --verb=get,list,watch --resource=pods |
| 113 | + kubectl create clusterrolebinding sdk-user-pod-creator --clusterrole=pod-creator --user=sdk-user |
| 114 | + kubectl create clusterrole service-reader --verb=get,list,watch --resource=services |
| 115 | + kubectl create clusterrolebinding sdk-user-service-reader --clusterrole=service-reader --user=sdk-user |
| 116 | + kubectl create clusterrole port-forward-pods --verb=create --resource=pods/portforward |
| 117 | + kubectl create clusterrolebinding sdk-user-port-forward-pods-binding --clusterrole=port-forward-pods --user=sdk-user |
| 118 | + kubectl config use-context sdk-user |
| 119 | +
|
| 120 | + - name: Run RayJob E2E tests |
| 121 | + run: | |
| 122 | + export CODEFLARE_TEST_OUTPUT_DIR=${{ env.TEMP_DIR }} |
| 123 | + echo "CODEFLARE_TEST_OUTPUT_DIR=${CODEFLARE_TEST_OUTPUT_DIR}" >> $GITHUB_ENV |
| 124 | +
|
| 125 | + set -euo pipefail |
| 126 | + pip install poetry |
| 127 | + poetry install --with test,docs |
| 128 | + echo "Running RayJob e2e tests..." |
| 129 | + poetry run pytest -v -s ./tests/e2e/rayjob/rayjob_lifecycled_cluster_test.py > ${CODEFLARE_TEST_OUTPUT_DIR}/pytest_output_rayjob.log 2>&1 |
| 130 | +
|
| 131 | + - name: Switch to kind-cluster context to print logs |
| 132 | + if: always() && steps.deploy.outcome == 'success' |
| 133 | + run: kubectl config use-context kind-cluster |
| 134 | + |
| 135 | + - name: Print Pytest output log |
| 136 | + if: always() && steps.deploy.outcome == 'success' |
| 137 | + run: | |
| 138 | + echo "Printing Pytest output logs" |
| 139 | + cat ${CODEFLARE_TEST_OUTPUT_DIR}/pytest_output_rayjob.log |
| 140 | +
|
| 141 | + - name: Print CodeFlare operator logs |
| 142 | + if: always() && steps.deploy.outcome == 'success' |
| 143 | + run: | |
| 144 | + echo "Printing CodeFlare operator logs" |
| 145 | + kubectl logs -n openshift-operators --tail -1 -l app.kubernetes.io/name=codeflare-operator | tee ${CODEFLARE_TEST_OUTPUT_DIR}/codeflare-operator.log |
| 146 | +
|
| 147 | + - name: Print KubeRay operator logs |
| 148 | + if: always() && steps.deploy.outcome == 'success' |
| 149 | + run: | |
| 150 | + echo "Printing KubeRay operator logs" |
| 151 | + kubectl logs -n default --tail -1 -l app.kubernetes.io/name=kuberay | tee ${CODEFLARE_TEST_OUTPUT_DIR}/kuberay.log |
| 152 | +
|
| 153 | + - name: Print Kueue controller logs |
| 154 | + if: always() && steps.deploy.outcome == 'success' |
| 155 | + run: | |
| 156 | + echo "Printing Kueue controller logs" |
| 157 | + kubectl logs -n kueue-system --tail -1 -l control-plane=controller-manager | tee ${CODEFLARE_TEST_OUTPUT_DIR}/kueue.log |
| 158 | +
|
| 159 | + - name: Export all KinD pod logs |
| 160 | + uses: ./common/github-actions/kind-export-logs |
| 161 | + if: always() && steps.deploy.outcome == 'success' |
| 162 | + with: |
| 163 | + output-directory: ${CODEFLARE_TEST_OUTPUT_DIR} |
| 164 | + |
| 165 | + - name: Upload logs |
| 166 | + uses: actions/upload-artifact@v4 |
| 167 | + if: always() && steps.deploy.outcome == 'success' |
| 168 | + with: |
| 169 | + name: logs |
| 170 | + retention-days: 10 |
| 171 | + path: | |
| 172 | + ${{ env.CODEFLARE_TEST_OUTPUT_DIR }}/**/*.log |
0 commit comments