[bug] azure-ai-projects-1.1.0b4 Cloud evaluation I/O error.

[mluker]

• azure-ai-projects:
• azure-ai-projects-1.1.0b4:
• Mac/Linux:
• 3.13:

Describe the bug
When running cloud evaluations using azure-ai-projects-1.1.0b4 there seems to be a race condition where the follow error occurs randomly: evaluation for run ID <123-456-789>: (InternalError) I/O operation on closed file.

Additional context
Full instruction and sample repo can be found here

To Reproduce
Create multiple clould evaulations in a loop and they will randomly fail with the above error.

Expected behavior
Able to run multiple cloud evaluations without file I/O issues.

Screenshots

2025-10-24 14:27:09 | INFO     | xxx-xxx-xxx | dataMapping provided: {'query': '${data.query}', 'response': '${data.response}', 'context': '${data.context}'}
2025-10-24 14:27:09 | INFO     | xxx-xxx-xxx | dataMapping provided: {'query': '${data.query}', 'response': '${data.response}'}
2025-10-24 14:27:09 | INFO     | xxx-xxx-xxx | dataMapping provided: {'response': '${data.response}', 'ground_truth': '${data.ground_truth}'}
2025-10-24 14:27:09 | INFO     | xxx-xxx-xxx | dataMapping provided: {'response': '${data.response}', 'ground_truth': '${data.ground_truth}'}
2025-10-24 14:27:09 | INFO     | xxx-xxx-xxx | Execute evaluators with evaluators: ['relevance', 'violence', 'f1_score', 'rouge']
======= Run Summary =======

Run name: "rouge_20251024_142709_646831"
Run status: "Completed"
Start time: "2025-10-24 14:27:09.646831+00:00"
Duration: "0:00:01.002103"

======= Run Summary =======

Run name: "f1_score_20251024_142709_647049"
Run status: "Completed"
Start time: "2025-10-24 14:27:09.647049+00:00"
Duration: "0:00:01.001694"

2025-10-24 14:27:40 | ERROR    | xxx-xxx-xxx | Error during evaluation for run ID xxx-xxx-xxx: (InternalError) I/O operation on closed file.
Traceback (most recent call last):
  File "/opt/venv/lib/python3.10/site-packages/azure/ai/evaluation/_evaluate/_evaluate.py", line 792, in evaluate
    return _evaluate(
  File "/opt/venv/lib/python3.10/site-packages/azure/ai/evaluation/_evaluate/_evaluate.py", line 925, in _evaluate
    eval_result_df, eval_metrics, per_evaluator_results = _run_callable_evaluators(
  File "/opt/venv/lib/python3.10/site-packages/azure/ai/evaluation/_evaluate/_evaluate.py", line 1193, in _run_callable_evaluators
    per_evaluator_results: Dict[str, __EvaluatorInfo] = {
  File "/opt/venv/lib/python3.10/site-packages/azure/ai/evaluation/_evaluate/_evaluate.py", line 1195, in <dictcomp>
    "result": batch_run_client.get_details(run, all_results=True),
  File "/opt/venv/lib/python3.10/site-packages/azure/ai/evaluation/_evaluate/_batch_run/_run_submitter_client.py", line 95, in get_details
    run = self._get_run(client_run)
  File "/opt/venv/lib/python3.10/site-packages/azure/ai/evaluation/_evaluate/_batch_run/_run_submitter_client.py", line 166, in _get_run
    return cast(Future[Run], run).result()
  File "/opt/miniconda/lib/python3.10/concurrent/futures/_base.py", line 458, in result
    return self.__get_result()
  File "/opt/miniconda/lib/python3.10/concurrent/futures/_base.py", line 403, in __get_result
    raise self._exception
  File "/opt/miniconda/lib/python3.10/concurrent/futures/thread.py", line 58, in run
    result = self.fn(*self.args, **self.kwargs)
  File "/opt/miniconda/lib/python3.10/asyncio/runners.py", line 44, in run
    return loop.run_until_complete(main)
  File "/opt/miniconda/lib/python3.10/asyncio/base_events.py", line 649, in run_until_complete
    return future.result()
  File "/opt/venv/lib/python3.10/site-packages/azure/ai/evaluation/_legacy/_batch_engine/_run_submitter.py", line 83, in submit
    self.stream_run(run=run, storage=local_storage, raise_on_error=self._config.raise_on_error)
  File "/opt/venv/lib/python3.10/site-packages/azure/ai/evaluation/_legacy/_batch_engine/_run_submitter.py", line 208, in stream_run
    RunSubmitter._print_run_summary(run, file_handler)
  File "/opt/venv/lib/python3.10/site-packages/azure/ai/evaluation/_legacy/_batch_engine/_run_submitter.py", line 243, in _print_run_summary
    text_out.write(
  File "/opt/venv/lib/python3.10/site-packages/azure/ai/evaluation/_legacy/_common/_logging.py", line 255, in write
    return self._prev_out.write(s)
  File "/opt/venv/lib/python3.10/site-packages/azure/ai/evaluation/_legacy/_common/_logging.py", line 255, in write
    return self._prev_out.write(s)
  File "/opt/venv/lib/python3.10/site-packages/azure/ai/evaluation/_legacy/_common/_logging.py", line 255, in write
    return self._prev_out.write(s)
ValueError: I/O operation on closed file.

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/app/evaluation/evaluation.py", line 44, in run_evaluation
    results = await execute_evaluators_async(evaluation, input_path, output_path, logger, log_path)
  File "/app/evaluation/eval_utils.py", line 500, in execute_evaluators_async
    return await asyncio.to_thread(
  File "/opt/miniconda/lib/python3.10/asyncio/threads.py", line 25, in to_thread
    return await loop.run_in_executor(None, func_call)
  File "/opt/miniconda/lib/python3.10/concurrent/futures/thread.py", line 58, in run
    result = self.fn(*self.args, **self.kwargs)
  File "/app/evaluation/eval_utils.py", line 486, in _execute_evaluators_blocking
    results = evaluate(
  File "/opt/venv/lib/python3.10/site-packages/azure/ai/evaluation/_evaluate/_evaluate.py", line 828, in evaluate
    raise EvaluationException(
azure.ai.evaluation._exceptions.EvaluationException: (InternalError) I/O operation on closed file.
2025-10-24 14:27:40 | INFO     | xxx-xxx-xxx | Start pending upload for evaluationResult
2025-10-24 14:27:40 | INFO     | xxx-xxx-xxx | start_upload_url: xxx
2025-10-24 14:27:40 | INFO     | xxx-xxx-xxx | start_upload_token_length: 2659, last 4 characters: uDAQ
2025-10-24 14:27:41 | INFO     | xxx-xxx-xxx | Successful response received with status: 200
2025-10-24 14:27:41 | INFO     | xxx-xxx-xxx | Uploading to Azure Storage as blob: /app/xxx/eval_results/user_logs.txt

[github-actions]

Thanks for the feedback! We are routing this to the appropriate team for follow-up. cc @dargilco @glharper @howieleung @jhakulin @nick863 @trangevi.

[dargilco]

Thank you @mluker for creating this issue and the detailed repro information! assigning to @singankit to triage.