vllm - 💡(How to fix) Fix [CI Failure]: Plugin Tests (2 GPUs) [1 participants]

[2026-04-17T04:39:06Z] ret = requests.post( [2026-04-17T04:39:06Z] server.url_for("pooling"), [2026-04-17T04:39:06Z] json=request_payload_url, [2026-04-17T04:39:06Z] ) [2026-04-17T04:39:06Z] [2026-04-17T04:39:06Z] response = ret.json() [2026-04-17T04:39:06Z] [2026-04-17T04:39:06Z] # verify the request response is in the correct format [2026-04-17T04:39:06Z] > assert (parsed_response := IOProcessorResponse(**response)) [2026-04-17T04:39:06Z] E pydantic_core._pydantic_core.ValidationError: 1 validation error for IOProcessorResponse [2026-04-17T04:39:06Z] E data [2026-04-17T04:39:06Z] E Field required [type=missing, input_value={'error': {'message': "'I...am': None, 'code': 500}}, input_type=dict] [2026-04-17T04:39:06Z] E For further information visit https://errors.pydantic.dev/2.12/v/missing [2026-04-17T04:39:06Z] [2026-04-17T04:39:06Z] plugins_tests/test_terratorch_io_processor_plugins.py:92: ValidationError

TL;DR

The test failure is likely due to a validation error caused by a missing required field in the server response, which can be mitigated by checking the server response for errors before attempting to parse it.

Guidance

• Investigate the server response to determine why the required data field is missing and instead contains an error object with a 500 status code.
• Modify the test to check the server response for errors before attempting to parse it as an IOProcessorResponse.
• Verify that the server is correctly handling the request and returning the expected response format.
• Consider adding error handling to the test to account for potential server errors and provide more informative error messages.

Notes

The exact cause of the server error is unclear and requires further investigation, but it appears to be related to the server returning an error response instead of the expected data.

Recommendation

Apply workaround: Modify the test to check the server response for errors before attempting to parse it, to prevent validation errors and provide more informative error messages.