Merge pull request GoogleCloudPlatform#3255 from mr0re1/health_out

Comment out pluging of `gpu-test`
ighosh98 · Nov 15, 2024 · c6ee4b1 · c6ee4b1
2 parents f62f2bc + 684c035
commit c6ee4b1
Showing 1 changed file with 2 additions and 1 deletion.
diff --git a/examples/machine-learning/a3-megagpu-8g/slurm-a3mega-cluster.yaml b/examples/machine-learning/a3-megagpu-8g/slurm-a3mega-cluster.yaml
@@ -226,7 +226,8 @@ deployment_groups:
           chmod 0755 "${SLURM_ROOT}/scripts/rxdm"
           ln -s "${SLURM_ROOT}/scripts/rxdm" "${SLURM_ROOT}/partition-$(vars.a3mega_partition_name)-prolog_slurmd.d/rxdm.prolog_slurmd"
           ln -s "${SLURM_ROOT}/scripts/rxdm" "${SLURM_ROOT}/partition-$(vars.a3mega_partition_name)-epilog_slurmd.d/rxdm.epilog_slurmd"
-          ln -s "/slurm/scripts/tools/gpu-test" "${SLURM_ROOT}/partition-$(vars.a3mega_partition_name)-epilog_slurmd.d/gpu-test.epilog_slurmd"
+          # Uncomment the line below to enable epilog that will check health of GPUs and drain node if problem is detected.
+          # ln -s "/slurm/scripts/tools/gpu-test" "${SLURM_ROOT}/partition-$(vars.a3mega_partition_name)-epilog_slurmd.d/gpu-test.epilog_slurmd"
       - type: shell
         destination: reset_enroot.sh
         content: |