@@ -53,27 +53,32 @@ jobs:
5353 env :
5454 BASE_IMAGE : ${{ needs.build-mpi-operator-compatible-base.outputs.DOCKER_TAG_FINAL }}
5555 TEST_NAME : ${{ matrix.test }}
56- JOB_NAME : " nccl-test-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.test }}"
57- LAUNCHER_NAME : " nccl-test-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.test }}-launcher"
58- TOKEN_NAME : " nccl-test-${{ github.run_id }}-${{ github.run_attempt }}-${{ matrix.test }}-token"
5956
6057
6158 steps :
6259 - name : Check out the repository
6360 uses : actions/checkout@v4
6461
62+ - name : Modify variables
63+ id : var
64+ shell : bash
65+ run : |
66+ echo "JOB_NAME=${{ env.JOB_NAME}//_/-}" >> $GITHUB_OUTPUT
67+ echo "LAUNCHER_NAME=nccl-test-${{ env.JOB_NAME}//_/-}-launcher" >> $GITHUB_OUTPUT
68+ echo "TOKEN_NAME=nccl-test-${{ env.JOB_NAME}//_/-}-token" >> $GITHUB_OUTPUT
69+
6570 - name : GHCR login and store K8s secret
6671 uses : ./.github/actions/ghcr-login
6772 with :
6873 docker-username : ${{ github.repository_owner }}
6974 docker-password : ${{ secrets.GITHUB_TOKEN }}
70- token-name : ${{ env .TOKEN_NAME }}
75+ token-name : ${{ steps.var .TOKEN_NAME }}
7176 - name : Configure Kubernetes job
7277 shell : bash
7378 run : |
74- export JOB_NAME="${{ env .JOB_NAME }}"
75- export LAUNCHER_NAME="${{ env .LAUNCHER_NAME }}"
76- export TOKEN_NAME="${{ env .TOKEN_NAME }}"
79+ export JOB_NAME="${{ steps.var .JOB_NAME }}"
80+ export LAUNCHER_NAME="${{ steps.var .LAUNCHER_NAME }}"
81+ export TOKEN_NAME="${{ steps.var .TOKEN_NAME }}"
7782 export TEST_NAME="${{ env.TEST_NAME }}"
7883 export WORKER_NAME="${JOB_NAME}-worker"
7984
@@ -95,11 +100,11 @@ jobs:
95100 uses : ./.github/actions/submit-k8s-job
96101 with :
97102 job-config-file : .github/eks-workflow-files/mpi-nccl-test.yml
98- job-name : ${{ env .LAUNCHER_NAME }}
103+ job-name : ${{ steps.var .LAUNCHER_NAME }}
99104 - name : Retrieve Kubernetes job status
100105 shell : bash -exo pipefail
101106 run : |
102- LAUNCHER_NAME="${{ env .LAUNCHER_NAME }}"
107+ LAUNCHER_NAME="${{ steps.var .LAUNCHER_NAME }}"
103108 while readarray -d : -t status < <(kubectl get job/${LAUNCHER_NAME} -o 'jsonpath={.status.failed}:{.status.succeeded}'); do
104109 failure=${status[0]:-0}
105110 success=${status[1]:-0}
@@ -118,7 +123,7 @@ jobs:
118123 if : failure()
119124 shell : bash
120125 run : |
121- LAUNCHER_NAME="${{ env .LAUNCHER_NAME }}"
126+ LAUNCHER_NAME="${{ steps.var .LAUNCHER_NAME }}"
122127 pods=$(kubectl get pods --selector=batch.kubernetes.io/job-name=${LAUNCHER_NAME} -o name)
123128 if [[ -n "${pods}" ]]; then
124129 kubectl describe ${pods}
@@ -127,9 +132,9 @@ jobs:
127132 if : always()
128133 uses : ./.github/actions/delete-k8s-job
129134 with :
130- job-name : ${{ env .LAUNCHER_NAME }}
135+ job-name : ${{ steps.var .LAUNCHER_NAME }}
131136 - name : Delete GitHub Container Registry token
132137 uses : ./.github/actions/delete-ghcr-token
133138 if : always()
134139 with :
135- token-name : ${{ env .TOKEN_NAME }}
140+ token-name : ${{ steps.var .TOKEN_NAME }}
0 commit comments