fix

Signed-off-by: Hemil Desai <[email protected]>

Author: hemildesai

nemo_run/run/ray/templates/ray_enroot.sub.j2

@@ -473,6 +473,11 @@ else
 # Args 1-N launch on worker nodes (nodes 1 through N-1)
 # Optional: set COMMAND='...' to run non-interactively instead of opening an interactive shell
 
+# Helper to get container PID
+get_container_pid() {
+  local node=\$1
+  srun --overlap --nodes=1 -w "\$node" --jobid $SLURM_JOB_ID bash -c "enroot list -f | awk 'NR>1 && \\\$2 ~ /^[0-9]+\\\$/ {print \\\$2; exit}'"
+}
 
 WORKER_NUM=\\\${1:-}
 if [[ -z "\\\$WORKER_NUM" ]]; then

test/core/execution/artifacts/expected_ray_cluster_enroot.sub

@@ -449,6 +449,11 @@ else
 # Args 1-N launch on worker nodes (nodes 1 through N-1)
 # Optional: set COMMAND='...' to run non-interactively instead of opening an interactive shell
 
+# Helper to get container PID
+get_container_pid() {
+  local node=\$1
+  srun --overlap --nodes=1 -w "\$node" --jobid $SLURM_JOB_ID bash -c "enroot list -f | awk 'NR>1 && \\\$2 ~ /^[0-9]+\\\$/ {print \\\$2; exit}'"
+}
 
 WORKER_NUM=\\\${1:-}
 if [[ -z "\\\$WORKER_NUM" ]]; then
@@ -482,4 +487,4 @@ EOF
   echo "     bash $CLUSTER_DIR/scripts/${SLURM_JOB_ID}-attach.sh 1  # to attach to worker 1"
   echo "     bash $CLUSTER_DIR/scripts/${SLURM_JOB_ID}-attach.sh 2  # to attach to worker 2, etc."
   sleep infinity
-fi
+fi