scripts/queue_scale_followup_after_pids.sh

#!/usr/bin/env bash
set -euo pipefail

cd "${PCMT_REMOTE_DIR:-/workspace/pcmt}"
source /venv/main/bin/activate || true
mkdir -p runs

for wait_pid in "$@"; do
  if [ -z "${wait_pid}" ]; then
    continue
  fi
  while kill -0 "${wait_pid}" 2>/dev/null; do
    echo "[scale-followup] waiting for pid=${wait_pid}"
    sleep 60
  done
done

bash scripts/run_scale_followup_gpu.sh