This commit is contained in:
Eugene Rakhmatulin
2026-04-15 22:32:14 -07:00
parent 76fbf0d0be
commit 6b7f8dace6

View File

@@ -977,8 +977,10 @@ exec_no_ray_cluster() {
worker_cmd="$clean --nnodes $total_nodes --node-rank $rank --master-addr $HEAD_IP --master-port $MASTER_PORT --headless" worker_cmd="$clean --nnodes $total_nodes --node-rank $rank --master-addr $HEAD_IP --master-port $MASTER_PORT --headless"
fi fi
echo "Launching worker (rank $rank) on $worker..." echo "Launching worker (rank $rank) on $worker..."
ssh -o BatchMode=yes -o StrictHostKeyChecking=no "$worker" \ local remote_payload remote_cmd
"docker exec -d $CONTAINER_NAME bash -c \"$worker_cmd >> /proc/1/fd/1 2>&1\"" remote_payload="$worker_cmd >> /proc/1/fd/1 2>&1"
printf -v remote_cmd 'docker exec -d %q bash -c %q' "$CONTAINER_NAME" "$remote_payload"
ssh -o BatchMode=yes -o StrictHostKeyChecking=no "$worker" "$remote_cmd"
(( rank++ )) (( rank++ ))
done done