Sporadically seeing "Error in _play_task" from Background Audio player

[danielgrittner]

Hey, we see the following error every now and then:

2025-10-08 15:30:30,524 - ERROR livekit.agents - Error in _play_task
Traceback (most recent call last):
  File "/.venv/lib/python3.12/site-packages/livekit/agents/utils/log.py", line 16, in async_fn_logs
    return await fn(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/.venv/lib/python3.12/site-packages/livekit/agents/voice/background_audio.py", line 367, in _play_task
    await gen.aclose()
RuntimeError: aclose(): asynchronous generator is already running

2025-10-08 15:30:30,526 - ERROR asyncio - Task exception was never retrieved
future: <Task finished name='Task-228' coro=<BackgroundAudioPlayer._play_task() done, defined at /.venv/lib/python3.12/site-packages/livekit/agents/utils/log.py:13> exception=RuntimeError('aclose(): asynchronous generator is already running')>
Traceback (most recent call last):
  File "/.venv/lib/python3.12/site-packages/livekit/agents/utils/log.py", line 16, in async_fn_logs
    return await fn(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/.venv/lib/python3.12/site-packages/livekit/agents/voice/background_audio.py", line 367, in _play_task
    await gen.aclose()
RuntimeError: aclose(): asynchronous generator is already running

We are on the newest version 1.2.14.

Unfortunately, I don't know how to reproduce this. I also found some old issues with the same problem stating it is supposed to be fixed. However, for us this is not the case.
I am observing this issue locally as well as in production.

[longcw]

previous attempt for fixing this issue #2814

I can still reproduce it when stop the PlayHandle very fast cc @theomonnom

It seems to be the race condition between await gen.aclose() and stream.__anext__() in audio mixer. I tried to ignore the RuntimeError of gen.aclose and add a break to the _gen_wrapper when play handle stopped, but the job process cannot fully exit if it breaks from the _gen_wrapper somehow...

import asyncio
import logging

from dotenv import load_dotenv

from livekit.agents import (
    AudioConfig,
    BackgroundAudioPlayer,
    BuiltinAudioClip,
    JobContext,
    WorkerOptions,
    cli,
)

logger = logging.getLogger("background-audio")

load_dotenv()


async def entrypoint(ctx: JobContext):
    await ctx.connect()

    background_audio = BackgroundAudioPlayer(
        ambient_sound=AudioConfig(BuiltinAudioClip.OFFICE_AMBIENCE, volume=1.0),
    )

    iterations = 20
    await background_audio.start(room=ctx.room)
    try:
        for i in range(iterations):
            logger.info(f"repeat {i}...")

            handle = background_audio.play(
                AudioConfig(BuiltinAudioClip.KEYBOARD_TYPING, volume=0.8)
            )

            # close it fast
            await asyncio.sleep(0.01)

            handle.stop()

            await asyncio.sleep(0.1)
    finally:
        await background_audio.aclose()

    logger.info("Test Done")


if __name__ == "__main__":
    cli.run_app(WorkerOptions(entrypoint_fnc=entrypoint))

[marioplumbarius]

I'm experiencing the same behavior. It doesn't seem to affect the agents, but it spams our logs.

[Timozegu]

Same thing here !

[berk-karahan-crea]

I’m using version 1.2.17 and encounter this error frequently. When it occurs, I notice delays in the agent’s responses and the background audio stops playing.

{"message": "Error in _play_task
Traceback (most recent call last):
  File \"/usr/local/lib/python3.12/site-packages/livekit/agents/utils/log.py\", line 16, in async_fn_logs
    return await fn(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^
  File \"/usr/local/lib/python3.12/site-packages/livekit/agents/voice/background_audio.py\", line 367, in _play_task
    await gen.aclose()
RuntimeError: aclose(): asynchronous generator is already running", "level": "ERROR",}

{"message": "Task exception was never retrieved
future: <Task finished name='Task-904' coro=<BackgroundAudioPlayer._play_task() done, defined at /usr/local/lib/python3.12/site-packages/livekit/agents/utils/log.py:13> exception=RuntimeError('aclose(): asynchronous generator is already running')>
Traceback (most recent call last):
  File \"/usr/local/lib/python3.12/site-packages/livekit/agents/utils/log.py\", line 16, in async_fn_logs
    return await fn(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^
  File \"/usr/local/lib/python3.12/site-packages/livekit/agents/voice/background_audio.py\", line 367, in _play_task
    await gen.aclose()
RuntimeError: aclose(): asynchronous generator is already running", "level": "ERROR",}

[davidzhao]

is anyone still seeing this on the latest 1.3.6?

[longcw]

@davidzhao I think it's still reproducible with the example above in #3606 (comment), it's a race condition when gen.aclose() is called during stream.__anext__().

[longcw]

It seems to be the race condition between await gen.aclose() and stream.anext() in audio mixer. I tried to ignore the RuntimeError of gen.aclose and add a break to the _gen_wrapper when play handle stopped, but the job process cannot fully exit if it breaks from the _gen_wrapper somehow...

found the reason why the process cannot exit: the AudioStreamDecoder of audio_frames_from_file is not closed properly, so the ThreadPoolExecutor created inside the decoder is not closed.