[Bug]: vLLM inference with direct runner leaves a python process in the background after Beam program terminates

[tvalentyn]

What happened?

I am consistently seeing a leftover process after running a pipeline similar to https://github.com/apache/beam/blob/master/examples/notebooks/beam-ml/run_inference_vllm.ipynb

root@759ad9652265:/workspace/vllm# py-spy dump --pid 5561
Process 5561: /usr/local/bin/python -c from multiprocessing.spawn import spawn_main; spawn_main(tracker_fd=17, pipe_handle=19) --multiprocessing-fork
Python v3.10.15 (/usr/local/bin/python3.10)

Thread 5561 (idle): "MainThread"
    poll (zmq/sugar/poll.py:106)
    poll (zmq/sugar/socket.py:1045)
    run_engine_loop (vllm/engine/multiprocessing/engine.py:192)
    start (vllm/engine/multiprocessing/engine.py:140)
    run_mp_engine (vllm/engine/multiprocessing/engine.py:370)
    run (multiprocessing/process.py:108)
    _bootstrap (multiprocessing/process.py:314)
    _main (multiprocessing/spawn.py:129)
    spawn_main (multiprocessing/spawn.py:116)
    <module> (<string>:1)
Thread 5838 (idle): "Thread-1"
    wait (threading.py:324)
    wait (threading.py:607)
    run (tqdm/_monitor.py:60)
    _bootstrap_inner (threading.py:1016)
    _bootstrap (threading.py:973)
Thread 5839 (idle): "Thread-2"
    wait (threading.py:324)
    wait (threading.py:607)
    run (tqdm/_monitor.py:60)
    _bootstrap_inner (threading.py:1016)
    _bootstrap (threading.py:973)
Thread 5896 (idle): "Thread-3"
    wait (threading.py:324)
    wait (threading.py:607)
    run (tqdm/_monitor.py:60)
    _bootstrap_inner (threading.py:1016)
    _bootstrap (threading.py:973)
Thread 6166 (idle): "Thread-5 (_report_usage_worker)"
    _report_continous_usage (vllm/usage/usage_lib.py:197)
    _report_usage_worker (vllm/usage/usage_lib.py:149)
    run (threading.py:953)
    _bootstrap_inner (threading.py:1016)
    _bootstrap (threading.py:973)

Issue Priority

Priority: 2 (default / most bugs should be filed as P2)

Issue Components

• Component: Python SDK
• Component: Java SDK
• Component: Go SDK
• Component: Typescript SDK
• Component: IO connector
• Component: Beam YAML
• Component: Beam examples
• Component: Beam playground
• Component: Beam katas
• Component: Website
• Component: Infrastructure
• Component: Spark Runner
• Component: Flink Runner
• Component: Samza Runner
• Component: Twister2 Runner
• Component: Hazelcast Jet Runner
• Component: Google Cloud Dataflow Runner

[tvalentyn]

cc: @damccorm fyi