harry0703/MoneyPrinterTurbo · warning · HttpException

{request_id}: task is still running

Error message

{request_id}: task is still running

What it means

Returned by DELETE /api/v1/tasks/{task_id} in app/controllers/v1/video.py when the task exists but tm.is_task_busy(task) reports it running (state or cross_post_state active). Deletion is refused with 409 because removing the record and directory out from under an executing thread would corrupt in-flight work. The server logs task_id, state, and cross_post_state at warning level for diagnosis.

Source

Thrown at app/controllers/v1/video.py:290

    )


@router.delete(
    "/tasks/{task_id}",
    response_model=TaskDeletionResponse,
    summary="Delete a generated short video task",
)
def delete_video(request: Request, task_id: str = Path(..., description="Task ID")):
    request_id = base.get_task_id(request)
    task = sm.state.get_task(task_id)
    if task:
        if tm.is_task_busy(task):
            logger.warning(
                f"refuse to delete busy task, request_id: {request_id}, "
                f"task_id: {task_id}, state: {task.get('state')}, "
                f"cross_post_state: {task.get('cross_post_state')}"
            )
            raise HttpException(
                task_id=task_id,
                status_code=409,
                message=f"{request_id}: task is still running",
            )

        tasks_dir = utils.task_dir()
        current_task_dir = os.path.join(tasks_dir, task_id)
        if os.path.exists(current_task_dir):
            shutil.rmtree(current_task_dir)

        sm.state.delete_task(task_id)
        logger.success(f"video deleted: {utils.to_json(task)}")
        return utils.get_response(200)

    raise HttpException(
        task_id=task_id, status_code=404, message=f"{request_id}: task not found"
    )

View on GitHub (pinned to 1f9f19c202)

Solutions

1. Poll GET /tasks/{id} until state reaches a terminal value, then retry the DELETE.
2. If the task appears wedged (busy far beyond normal duration), investigate the worker/queue state before force-cleaning directories manually.
3. Make cleanup scripts skip tasks whose state is running/queued rather than deleting blindly.

Example fix

# before
requests.delete(f"{base}/api/v1/tasks/{task_id}", headers=h).raise_for_status()

# after
import time
while True:
    task = requests.get(f"{base}/api/v1/tasks/{task_id}", headers=h).json()["data"]
    if task["state"] not in ("running", "queued", "pending"):
        break
    time.sleep(5)
requests.delete(f"{base}/api/v1/tasks/{task_id}", headers=h).raise_for_status()

task = requests.get(f"{base}/api/v1/tasks/{task_id}", headers=h).json()["data"]
if task.get("state") in BUSY_STATES or task.get("cross_post_state") in BUSY_STATES:
    # defer the delete until terminal
    schedule_retry_later(task_id)

BUSY_STATES = {"running", "queued", "pending", "processing"}
def is_task_idle(task: dict) -> bool:
    return task.get("state") not in BUSY_STATES and task.get("cross_post_state") not in BUSY_STATES

for _ in range(MAX_WAIT // POLL_SEC):
    resp = requests.delete(f"{base}/api/v1/tasks/{task_id}", headers=h)
    if resp.status_code == 409:
        time.sleep(POLL_SEC); continue
    if resp.status_code == 404:
        return  # already gone
    resp.raise_for_status(); return
raise RuntimeError("task stayed busy; manual intervention needed")

When it happens

Trigger: DELETE while the task state is a running/queued state; deleting immediately after POST /videos without waiting for completion; deleting while a cross-post step is still in flight.

Common situations: Cleanup scripts that delete 'all tasks' including in-flight ones; UI delete buttons enabled before the status terminal event; long generations where the user assumes completion but cross_post_state is still active.

Related errors

• {request_id}: task not found

AI-assisted analysis of harry0703/MoneyPrinterTurbo@1f9f19c202 (2026-08-14). Data as JSON: /api/errors/0a2c0fdf38e51c23. Report an issue: GitHub.