Random error thrown on response #88

MarkusPfundstein · 2024-12-04T09:45:48Z

Describe the bug
Sometimes, I see a stacktrace printed in the logs of my mcp server. Claude eventually succeeds to response but I think its good to investigate it.

To Reproduce
Its hard to reproduce as it does not always happen. The code in my codebase that caused it to happen is this one:

    def run_tool(self, args: dict) -> Sequence[TextContent | ImageContent | EmbeddedResource]:

        user_id = args.get(toolhandler.USER_ID_ARG)
        if not user_id:
            raise RuntimeError(f"Missing required argument: {toolhandler.USER_ID_ARG}")

        gmail_service = gmail.GmailService(user_id=user_id)
        query = args.get('query')
        max_results = args.get('max_results', 100)
        emails = gmail_service.query_emails(query=query, max_results=max_results)

        return [
            TextContent(
                type="text",
                text=json.dumps(emails, indent=2)
            )
        ]

Called in this context

@app.call_tool()
async def call_tool(name: str, arguments: Any) -> Sequence[TextContent | ImageContent | EmbeddedResource]:
    """Handle tool calls for command line run."""
    
    if not isinstance(arguments, dict):
        raise RuntimeError("arguments must be dictionary")
    
    if toolhandler.USER_ID_ARG not in arguments:
        raise RuntimeError("user_id argument is missing in dictionary.")

    setup_oauth2(user_id=arguments.get(toolhandler.USER_ID_ARG, ""))

    tool_handler = get_tool_handler(name)
    if not tool_handler:
        raise ValueError(f"Unknown tool: {name}")

    try:
        return tool_handler.run_tool(arguments)
    except Exception as e:
        logger.error(str(e))
        raise RuntimeError(f"Caught Exception. Error: {str(e)}")

The RunTime Error is NOT thrown

Expected behavior
No stacktrace?

Desktop (please complete the following information):

MAC OSX
Python 3.13
mcp 1.1.0

Additional context

Here is a full log

+ Exception Group Traceback (most recent call last):
  |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/mcp/server/stdio.py", line 83, in stdio_server
  |     yield read_stream, write_stream
  |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/src/mcp_gsuite/server.py", line 161, in main
  |     await app.run(
  |     ...<3 lines>...
  |     )
  |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/mcp/server/__init__.py", line 423, in run
  |     async with ServerSession(
  |                ~~~~~~~~~~~~~^
  |         read_stream, write_stream, initialization_options
  |         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  |     ) as session:
  |     ^
  |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/mcp/shared/session.py", line 122, in __aexit__
  |     return await self._task_group.__aexit__(exc_type, exc_val, exc_tb)
  |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/anyio/_backends/_asyncio.py", line 763, in __aexit__
  |     raise BaseExceptionGroup(
  |         "unhandled errors in a TaskGroup", self._exceptions
  |     )
  | ExceptionGroup: unhandled errors in a TaskGroup (1 sub-exception)
  +-+---------------- 1 ----------------
    | Traceback (most recent call last):
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/mcp/shared/session.py", line 235, in _receive_loop
    |     notification = self._receive_notification_type.model_validate(
    |         message.root.model_dump(
    |             by_alias=True, mode="json", exclude_none=True
    |         )
    |     )
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/pydantic/main.py", line 627, in model_validate
    |     return cls.__pydantic_validator__.validate_python(
    |            ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^
    |         obj, strict=strict, from_attributes=from_attributes, context=context
    |         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    |     )
    |     ^
    | pydantic_core._pydantic_core.ValidationError: 5 validation errors for ClientNotification
    | ProgressNotification.method
    |   Input should be 'notifications/progress' [type=literal_error, input_value='cancelled', input_type=str]
    |     For further information visit https://errors.pydantic.dev/2.10/v/literal_error
    | ProgressNotification.params.progressToken
    |   Field required [type=missing, input_value={'requestId': 20, 'reason... -2: Request timed out'}, input_type=dict]
    |     For further information visit https://errors.pydantic.dev/2.10/v/missing
    | ProgressNotification.params.progress
    |   Field required [type=missing, input_value={'requestId': 20, 'reason... -2: Request timed out'}, input_type=dict]
    |     For further information visit https://errors.pydantic.dev/2.10/v/missing
    | InitializedNotification.method
    |   Input should be 'notifications/initialized' [type=literal_error, input_value='cancelled', input_type=str]
    |     For further information visit https://errors.pydantic.dev/2.10/v/literal_error
    | RootsListChangedNotification.method
    |   Input should be 'notifications/roots/list_changed' [type=literal_error, input_value='cancelled', input_type=str]
    |     For further information visit https://errors.pydantic.dev/2.10/v/literal_error
    +------------------------------------

During handling of the above exception, another exception occurred:

  + Exception Group Traceback (most recent call last):
  |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/bin/mcp-gsuite", line 8, in <module>
  |     sys.exit(main())
  |              ~~~~^^
  |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/src/mcp_gsuite/__init__.py", line 6, in main
  |     asyncio.run(server.main())
  |     ~~~~~~~~~~~^^^^^^^^^^^^^^^
  |   File "/opt/homebrew/Cellar/[email protected]/3.13.0_1/Frameworks/Python.framework/Versions/3.13/lib/python3.13/asyncio/runners.py", line 194, in run
  |     return runner.run(main)
  |            ~~~~~~~~~~^^^^^^
  |   File "/opt/homebrew/Cellar/[email protected]/3.13.0_1/Frameworks/Python.framework/Versions/3.13/lib/python3.13/asyncio/runners.py", line 118, in run
  |     return self._loop.run_until_complete(task)
  |            ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^
  |   File "/opt/homebrew/Cellar/[email protected]/3.13.0_1/Frameworks/Python.framework/Versions/3.13/lib/python3.13/asyncio/base_events.py", line 721, in run_until_complete
  |     return future.result()
  |            ~~~~~~~~~~~~~^^
  |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/src/mcp_gsuite/server.py", line 160, in main
  |     async with stdio_server() as (read_stream, write_stream):
  |                ~~~~~~~~~~~~^^
  |   File "/opt/homebrew/Cellar/[email protected]/3.13.0_1/Frameworks/Python.framework/Versions/3.13/lib/python3.13/contextlib.py", line 235, in __aexit__
  |     await self.gen.athrow(value)
  |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/mcp/server/stdio.py", line 80, in stdio_server
  |     async with anyio.create_task_group() as tg:
  |                ~~~~~~~~~~~~~~~~~~~~~~~^^
  |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/anyio/_backends/_asyncio.py", line 763, in __aexit__
  |     raise BaseExceptionGroup(
  |         "unhandled errors in a TaskGroup", self._exceptions
  |     )
  | ExceptionGroup: unhandled errors in a TaskGroup (1 sub-exception)
  +-+---------------- 1 ----------------
    | Traceback (most recent call last):
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/anyio/streams/memory.py", line 242, in send
    |     self.send_nowait(item)
    |     ~~~~~~~~~~~~~~~~^^^^^^
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/anyio/streams/memory.py", line 225, in send_nowait
    |     raise WouldBlock
    | anyio.WouldBlock
    | 
    | During handling of the above exception, another exception occurred:
    | 
    | Traceback (most recent call last):
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/mcp/server/__init__.py", line 467, in run
    |     await message.respond(
    |     ...<4 lines>...
    |     )
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/mcp/shared/session.py", line 58, in respond
    |     await self._session._send_response(
    |         request_id=self.request_id, response=response
    |     )
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/mcp/shared/session.py", line 196, in _send_response
    |     await self._write_stream.send(JSONRPCMessage(jsonrpc_error))
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/anyio/streams/memory.py", line 248, in send
    |     await send_event.wait()
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/anyio/_backends/_asyncio.py", line 1747, in wait
    |     await self._event.wait()
    |   File "/opt/homebrew/Cellar/[email protected]/3.13.0_1/Frameworks/Python.framework/Versions/3.13/lib/python3.13/asyncio/locks.py", line 213, in wait
    |     await fut
    | asyncio.exceptions.CancelledError: Cancelled by cancel scope 1040d7c50
    | 
    | During handling of the above exception, another exception occurred:
    | 
    | Exception Group Traceback (most recent call last):
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/mcp/server/stdio.py", line 83, in stdio_server
    |     yield read_stream, write_stream
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/src/mcp_gsuite/server.py", line 161, in main
    |     await app.run(
    |     ...<3 lines>...
    |     )
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/mcp/server/__init__.py", line 423, in run
    |     async with ServerSession(
    |                ~~~~~~~~~~~~~^
    |         read_stream, write_stream, initialization_options
    |         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    |     ) as session:
    |     ^
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/mcp/shared/session.py", line 122, in __aexit__
    |     return await self._task_group.__aexit__(exc_type, exc_val, exc_tb)
    |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/anyio/_backends/_asyncio.py", line 763, in __aexit__
    |     raise BaseExceptionGroup(
    |         "unhandled errors in a TaskGroup", self._exceptions
    |     )
    | ExceptionGroup: unhandled errors in a TaskGroup (1 sub-exception)
    +-+---------------- 1 ----------------
      | Traceback (most recent call last):
      |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/mcp/shared/session.py", line 235, in _receive_loop
      |     notification = self._receive_notification_type.model_validate(
      |         message.root.model_dump(
      |             by_alias=True, mode="json", exclude_none=True
      |         )
      |     )
      |   File "/Users/markus/experiments/claude-mvp/mcp-gsuite/.venv/lib/python3.13/site-packages/pydantic/main.py", line 627, in model_validate
      |     return cls.__pydantic_validator__.validate_python(
      |            ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^
      |         obj, strict=strict, from_attributes=from_attributes, context=context
      |         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
      |     )
      |     ^
      | pydantic_core._pydantic_core.ValidationError: 5 validation errors for ClientNotification
      | ProgressNotification.method
      |   Input should be 'notifications/progress' [type=literal_error, input_value='cancelled', input_type=str]
      |     For further information visit https://errors.pydantic.dev/2.10/v/literal_error
      | ProgressNotification.params.progressToken
      |   Field required [type=missing, input_value={'requestId': 20, 'reason... -2: Request timed out'}, input_type=dict]
      |     For further information visit https://errors.pydantic.dev/2.10/v/missing
      | ProgressNotification.params.progress
      |   Field required [type=missing, input_value={'requestId': 20, 'reason... -2: Request timed out'}, input_type=dict]
      |     For further information visit https://errors.pydantic.dev/2.10/v/missing
      | InitializedNotification.method
      |   Input should be 'notifications/initialized' [type=literal_error, input_value='cancelled', input_type=str]
      |     For further information visit https://errors.pydantic.dev/2.10/v/literal_error
      | RootsListChangedNotification.method
      |   Input should be 'notifications/roots/list_changed' [type=literal_error, input_value='cancelled', input_type=str]
      |     For further information visit https://errors.pydantic.dev/2.10/v/literal_error

The text was updated successfully, but these errors were encountered:

michail-nikolaev · 2024-12-04T14:33:41Z

Got the same on Windows:

During handling of the above exception, another exception occurred:

  + Exception Group Traceback (most recent call last):
  |   File "C:\Dev\mcp\servers\src\test-server\src\test_server\server.py", line 78, in <module>
  |     asyncio.run(serve())
  |   File "C:\Users\user\AppData\Local\Programs\Python\Python312\Lib\asyncio\runners.py", line 194, in run
  |     return runner.run(main)
  |            ^^^^^^^^^^^^^^^^
  |   File "C:\Users\user\AppData\Local\Programs\Python\Python312\Lib\asyncio\runners.py", line 118, in run
  |     return self._loop.run_until_complete(task)
  |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  |   File "C:\Users\user\AppData\Local\Programs\Python\Python312\Lib\asyncio\base_events.py", line 687, in run_until_complete
  |     return future.result()
  |            ^^^^^^^^^^^^^^^
  |   File "C:\Dev\mcp\servers\src\test-server\src\test_server\server.py", line 72, in serve
  |     async with stdio_server() as (read_stream, write_stream):
  |                ^^^^^^^^^^^^^^
  |   File "C:\Users\user\AppData\Local\Programs\Python\Python312\Lib\contextlib.py", line 231, in __aexit__
  |     await self.gen.athrow(value)
  |   File "C:\Dev\mcp\servers\src\git\.venv\Lib\site-packages\mcp\server\stdio.py", line 80, in stdio_server
  |     async with anyio.create_task_group() as tg:
  |                ^^^^^^^^^^^^^^^^^^^^^^^^^
  |   File "C:\Dev\mcp\servers\src\git\.venv\Lib\site-packages\anyio\_backends\_asyncio.py", line 763, in __aexit__
  |     raise BaseExceptionGroup(
  | ExceptionGroup: unhandled errors in a TaskGroup (1 sub-exception)
  +-+---------------- 1 ----------------
    | Exception Group Traceback (most recent call last):
    |   File "C:\Dev\mcp\servers\src\git\.venv\Lib\site-packages\mcp\server\stdio.py", line 83, in stdio_server
    |     yield read_stream, write_stream
    |   File "C:\Dev\mcp\servers\src\test-server\src\test_server\server.py", line 73, in serve
    |     await server.run(read_stream, write_stream, options, raise_exceptions=True)
    |   File "C:\Dev\mcp\servers\src\git\.venv\Lib\site-packages\mcp\server\__init__.py", line 404, in run
    |     async with ServerSession(
    |                ^^^^^^^^^^^^^^
    |   File "C:\Dev\mcp\servers\src\git\.venv\Lib\site-packages\mcp\shared\session.py", line 122, in __aexit__
    |     return await self._task_group.__aexit__(exc_type, exc_val, exc_tb)
    |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    |   File "C:\Dev\mcp\servers\src\git\.venv\Lib\site-packages\anyio\_backends\_asyncio.py", line 763, in __aexit__
    |     raise BaseExceptionGroup(
    | ExceptionGroup: unhandled errors in a TaskGroup (1 sub-exception)
    +-+---------------- 1 ----------------
      | Traceback (most recent call last):
      |   File "C:\Dev\mcp\servers\src\git\.venv\Lib\site-packages\mcp\shared\session.py", line 235, in _receive_loop
      |     notification = self._receive_notification_type.model_validate(
      |                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
      |   File "C:\Dev\mcp\servers\src\git\.venv\Lib\site-packages\pydantic\main.py", line 627, in model_validate
      |     return cls.__pydantic_validator__.validate_python(
      |            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
      | pydantic_core._pydantic_core.ValidationError: 5 validation errors for ClientNotification
ProgressNotification.method
  Input should be 'notifications/progress' [type=literal_error, input_value='cancelled', input_type=str]
    For further information visit https://errors.pydantic.dev/2.10/v/literal_error
ProgressNotification.params.progressToken
  Field required [type=missing, input_value={'requestId': 25, 'reason... -2: Request timed out'}, input_type=dict]
    For further information visit https://errors.pydantic.dev/2.10/v/missing
ProgressNotification.params.progress
  Field required [type=missing, input_value={'requestId': 25, 'reason... -2: Request timed out'}, input_type=dict]
    For further information visit https://errors.pydantic.dev/2.10/v/missing
InitializedNotification.method
  Input should be 'notifications/initialized' [type=literal_error, input_value='cancelled', input_type=str]
    For further information visit https://errors.pydantic.dev/2.10/v/literal_error
RootsListChangedNotification.method
  Input should be 'notifications/roots/list_changed' [type=literal_error, input_value='cancelled', input_type=str]
    For further information visit https://errors.pydantic.dev/2.10/v/literal_error
      +------------------------------------

lgesuellip · 2024-12-05T14:43:33Z

Hey team,

I am encountering the same error. When the server-side response is delayed due to an extended process (>10 seconds), the error occurs. Although the server successfully returns the results, it crashes afterward, blocking the next requests.
Is there any way to fix it? I am on a Mac, using Claude Desktop as a client.

7shi · 2024-12-06T06:29:49Z

I've investigated the issue reported by @MarkusPfundstein in the mcp-obsidian project (MarkusPfundstein/mcp-obsidian#3).

I was able to reproduce this issue with a minimal MCP server implementation that simply waits for a specified number of seconds:

https://github.com/7shi/mcp-wait/blob/main/py/server.py

The behavior varies depending on the client:

With MCP Inspector (10-second timeout):

Server crashes after sending results if processing takes 8-10 seconds or longer
The crash occurs regardless of the timeout setting
After crashing, the server stops responding to all requests

With Claude Desktop app (1-minute timeout):

Same behavior: server crashes after long-running operations (8-10 seconds or more)
The crash appears to be related to the processing duration rather than the timeout settings

This issue seems to be specific to the Python implementation of MCP servers. For comparison, I also implemented the same functionality in TypeScript, which continues to operate normally even after timeout events.

https://github.com/7shi/mcp-wait/blob/main/ts/src/index.ts

The crash occurs at the point where the server tries to send results back to the client after a long-running operation, making it particularly problematic for operations that involve external API calls or other time-consuming tasks.

Steps to reproduce:

Start MCP Inspector with the server:
npx @modelcontextprotocol/inspector uv run python server.py
Call wait_seconds with seconds=11
- The request will timeout
Try to call wait_seconds again
- The server has crashed and won't accept any further requests

The key characteristic of this issue is that the server successfully returns results for long-running operations, but crashes afterwards, preventing it from accepting any subsequent requests. This makes it particularly problematic for scenarios involving external API calls or other time-consuming tasks, as the server becomes unusable after the first long operation.

lgesuellip · 2024-12-06T21:35:19Z

@7shi I am facing the same errors. Have you fixed them in some way?
Have you tried using SSE?

7shi · 2024-12-07T04:36:43Z

@lgesuellip I don't know how to fix it. I have not tried SSE.

dsp-ant · 2024-12-10T14:11:30Z

Thank you for the repro. That is very useful. That is helpful. I'll take a look. Can't promise any dates.

dsp-ant added the bug Something isn't working label Dec 4, 2024

dsp-ant self-assigned this Dec 4, 2024

7shi mentioned this issue Dec 6, 2024

Obsidian Rest API starts hanging forever when patch content is called with invalid target MarkusPfundstein/mcp-obsidian#3

Closed

rusiaaman mentioned this issue Dec 15, 2024

Added cancelled handler. Fixes #88 #102

Closed

9 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Random error thrown on response #88

Random error thrown on response #88

MarkusPfundstein commented Dec 4, 2024

michail-nikolaev commented Dec 4, 2024

lgesuellip commented Dec 5, 2024 •

edited

Loading

7shi commented Dec 6, 2024

lgesuellip commented Dec 6, 2024

7shi commented Dec 7, 2024

dsp-ant commented Dec 10, 2024

Random error thrown on response #88

Random error thrown on response #88

Comments

MarkusPfundstein commented Dec 4, 2024

michail-nikolaev commented Dec 4, 2024

lgesuellip commented Dec 5, 2024 • edited Loading

7shi commented Dec 6, 2024

lgesuellip commented Dec 6, 2024

7shi commented Dec 7, 2024

dsp-ant commented Dec 10, 2024

lgesuellip commented Dec 5, 2024 •

edited

Loading