Fix: MCP cannot handle empty Auth field properly (#11034)

### What problem does this PR solve?

Fix MCP cannot handle empty Auth field properly, then result in 

```bash
2025-11-05 11:10:41,919 INFO     51209 Negotiated protocol version: 2025-06-18
2025-11-05 11:10:41,920 INFO     51209 client_session initialized successfully
2025-11-05 11:10:41,994 INFO     51209 127.0.0.1 - - [05/Nov/2025 11:10:41] "GET /api/v1/datasets?page=1&page_size=1000&orderby=create_time&desc=True HTTP/1.1" 200 -
2025-11-05 11:10:41,999 INFO     51209 Want to clean up 1 MCP sessions
2025-11-05 11:10:42,000 INFO     51209 1 MCP sessions has been cleaned up. 0 in global context.
2025-11-05 11:10:42,001 INFO     51209 127.0.0.1 - - [05/Nov/2025 11:10:42] "POST /v1/mcp_server/test_mcp HTTP/1.1" 200 -
2025-11-05 11:11:30,441 INFO     51209 Negotiated protocol version: 2025-06-18
2025-11-05 11:11:30,442 INFO     51209 client_session initialized successfully
2025-11-05 11:11:30,520 INFO     51209 127.0.0.1 - - [05/Nov/2025 11:11:30] "GET /api/v1/datasets?page=1&page_size=1000&orderby=create_time&desc=True HTTP/1.1" 200 -
2025-11-05 11:11:30,525 INFO     51209 Want to clean up 1 MCP sessions
2025-11-05 11:11:30,526 INFO     51209 1 MCP sessions has been cleaned up. 0 in global context.
2025-11-05 11:11:30,527 INFO     51209 127.0.0.1 - - [05/Nov/2025 11:11:30] "POST /v1/mcp_server/test_mcp HTTP/1.1" 200 -
2025-11-05 11:11:31,476 INFO     51209 Negotiated protocol version: 2025-06-18
2025-11-05 11:11:31,476 INFO     51209 client_session initialized successfully
2025-11-05 11:11:31,549 INFO     51209 127.0.0.1 - - [05/Nov/2025 11:11:31] "GET /api/v1/datasets?page=1&page_size=1000&orderby=create_time&desc=True HTTP/1.1" 200 -
2025-11-05 11:11:31,552 INFO     51209 Want to clean up 1 MCP sessions
2025-11-05 11:11:31,553 INFO     51209 1 MCP sessions has been cleaned up. 0 in global context.
2025-11-05 11:11:31,554 INFO     51209 127.0.0.1 - - [05/Nov/2025 11:11:31] "POST /v1/mcp_server/test_mcp HTTP/1.1" 200 -
2025-11-05 11:11:51,930 ERROR    51209 unhandled errors in a TaskGroup (1 sub-exception)
  + Exception Group Traceback (most recent call last):
  |   File "/home/xxxxxxxxx/workspace/ragflow/rag/utils/mcp_tool_call_conn.py", line 86, in _mcp_server_loop
  |     async with streamablehttp_client(url, headers) as (read_stream, write_stream, _):
  |   File "/home/xxxxxxxxx/.local/share/uv/python/cpython-3.10.16-linux-x86_64-gnu/lib/python3.10/contextlib.py", line 217, in __aexit__
  |     await self.gen.athrow(typ, value, traceback)
  |   File "/home/xxxxxxxxx/workspace/ragflow/.venv/lib/python3.10/site-packages/mcp/client/streamable_http.py", line 478, in streamablehttp_client
  |     async with anyio.create_task_group() as tg:
  |   File "/home/xxxxxxxxx/workspace/ragflow/.venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 781, in __aexit__
  |     raise BaseExceptionGroup(
  | exceptiongroup.ExceptionGroup: unhandled errors in a TaskGroup (1 sub-exception)
  +-+---------------- 1 ----------------
    | Traceback (most recent call last):
    |   File "/home/xxxxxxxxx/workspace/ragflow/.venv/lib/python3.10/site-packages/mcp/client/streamable_http.py", line 409, in handle_request_async
    |     await self._handle_post_request(ctx)
    |   File "/home/xxxxxxxxx/workspace/ragflow/.venv/lib/python3.10/site-packages/mcp/client/streamable_http.py", line 278, in _handle_post_request
    |     response.raise_for_status()
    |   File "/home/xxxxxxxxx/workspace/ragflow/.venv/lib/python3.10/site-packages/httpx/_models.py", line 829, in raise_for_status
    |     raise HTTPStatusError(message, request=request, response=self)
    | httpx.HTTPStatusError: Server error '502 Bad Gateway' for url 'http://192.168.1.38:9382/mcp'
    | For more information check: https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/502
    +------------------------------------
2025-11-05 11:11:51,942 ERROR    51209 Error fetching tools from MCP server: streamable-http: http://192.168.1.38:9382/mcp
Traceback (most recent call last):
  File "/home/xxxxxxxxx/workspace/ragflow/rag/utils/mcp_tool_call_conn.py", line 168, in get_tools
    return future.result(timeout=timeout)
  File "/home/xxxxxxxxx/.local/share/uv/python/cpython-3.10.16-linux-x86_64-gnu/lib/python3.10/concurrent/futures/_base.py", line 458, in result
    return self.__get_result()
  File "/home/xxxxxxxxx/.local/share/uv/python/cpython-3.10.16-linux-x86_64-gnu/lib/python3.10/concurrent/futures/_base.py", line 403, in __get_result
    raise self._exception
  File "<@beartype(rag.utils.mcp_tool_call_conn.MCPToolCallSession._get_tools_from_mcp_server) at 0x7d58f02e2c20>", line 40, in _get_tools_from_mcp_server
  File "/home/xxxxxxxxx/workspace/ragflow/rag/utils/mcp_tool_call_conn.py", line 160, in _get_tools_from_mcp_server
    result: ListToolsResult = await self._call_mcp_server("list_tools", timeout=timeout)
  File "<@beartype(rag.utils.mcp_tool_call_conn.MCPToolCallSession._call_mcp_server) at 0x7d58f02e2b00>", line 63, in _call_mcp_server
  File "/home/xxxxxxxxx/workspace/ragflow/rag/utils/mcp_tool_call_conn.py", line 139, in _call_mcp_server
    raise result
ValueError: Connection failed (possibly due to auth error). Please check authentication settings first
2025-11-05 11:11:51,943 ERROR    51209 Test MCP error: Connection failed (possibly due to auth error). Please check authentication settings first
Traceback (most recent call last):
  File "/home/xxxxxxxxx/workspace/ragflow/api/apps/mcp_server_app.py", line 429, in test_mcp
    tools = tool_call_session.get_tools(timeout)
  File "<@beartype(rag.utils.mcp_tool_call_conn.MCPToolCallSession.get_tools) at 0x7d58f02e2cb0>", line 40, in get_tools
  File "/home/xxxxxxxxx/workspace/ragflow/rag/utils/mcp_tool_call_conn.py", line 168, in get_tools
    return future.result(timeout=timeout)
  File "/home/xxxxxxxxx/.local/share/uv/python/cpython-3.10.16-linux-x86_64-gnu/lib/python3.10/concurrent/futures/_base.py", line 458, in result
    return self.__get_result()
  File "/home/xxxxxxxxx/.local/share/uv/python/cpython-3.10.16-linux-x86_64-gnu/lib/python3.10/concurrent/futures/_base.py", line 403, in __get_result
    raise self._exception
  File "<@beartype(rag.utils.mcp_tool_call_conn.MCPToolCallSession._get_tools_from_mcp_server) at 0x7d58f02e2c20>", line 40, in _get_tools_from_mcp_server
  File "/home/xxxxxxxxx/workspace/ragflow/rag/utils/mcp_tool_call_conn.py", line 160, in _get_tools_from_mcp_server
    result: ListToolsResult = await self._call_mcp_server("list_tools", timeout=timeout)
  File "<@beartype(rag.utils.mcp_tool_call_conn.MCPToolCallSession._call_mcp_server) at 0x7d58f02e2b00>", line 63, in _call_mcp_server
  File "/home/xxxxxxxxx/workspace/ragflow/rag/utils/mcp_tool_call_conn.py", line 139, in _call_mcp_server
    raise result
ValueError: Connection failed (possibly due to auth error). Please check authentication settings first
2025-11-05 11:11:51,944 INFO     51209 Want to clean up 1 MCP sessions
2025-11-05 11:11:51,945 INFO     51209 1 MCP sessions has been cleaned up. 0 in global context.
2025-11-05 11:11:51,946 INFO     51209 127.0.0.1 - - [05/Nov/2025 11:11:51] "POST /v1/mcp_server/test_mcp HTTP/1.1" 200 -
2025-11-05 11:12:20,484 INFO     51209 Negotiated protocol version: 2025-06-18
2025-11-05 11:12:20,485 INFO     51209 client_session initialized successfully
2025-11-05 11:12:20,570 INFO     51209 127.0.0.1 - - [05/Nov/2025 11:12:20] "GET /api/v1/datasets?page=1&page_size=1000&orderby=create_time&desc=True HTTP/1.1" 200 -
2025-11-05 11:12:20,573 INFO     51209 Want to clean up 1 MCP sessions
2025-11-05 11:12:20,574 INFO     51209 1 MCP sessions has been cleaned up. 0 in global context.
2025-11-05 11:12:20,575 INFO     51209 127.0.0.1 - - [05/Nov/2025 11:12:20] "POST /v1/mcp_server/test_mcp HTTP/1.1" 200 -
2025-11-05 11:15:02,119 INFO     51209 127.0.0.1 - - [05/Nov/2025 11:15:02] "GET /api/v1/datasets?page=1&page_size=1000&orderby=create_time&desc=True HTTP/1.1" 200 -
2025-11-05 11:16:24,967 INFO     51209 127.0.0.1 - - [05/Nov/2025 11:16:24] "GET /api/v1/datasets?page=1&page_size=1000&orderby=create_time&desc=True HTTP/1.1" 200 -
2025-11-05 11:30:24,284 ERROR    51209 Task was destroyed but it is pending!
task: <Task pending name='Task-58' coro=<MCPToolCallSession._mcp_server_loop() running at <@beartype(rag.utils.mcp_tool_call_conn.MCPToolCallSession._mcp_server_loop) at 0x7d58f02e29e0>:11> wait_for=<Future pending cb=[Task.task_wakeup()]> cb=[_chain_future.<locals>._call_set_state() at /home/xxxxxxxxx/.local/share/uv/python/cpython-3.10.16-linux-x86_64-gnu/lib/python3.10/asyncio/futures.py:392]>
2025-11-05 11:30:24,285 ERROR    51209 Task was destroyed but it is pending!
task: <Task pending name='Task-67' coro=<Queue.get() running at /home/xxxxxxxxx/.local/share/uv/python/cpython-3.10.16-linux-x86_64-gnu/lib/python3.10/asyncio/queues.py:159> wait_for=<Future pending cb=[Task.task_wakeup()]> cb=[_release_waiter(<Future pendi...ask_wakeup()]>)() at /home/xxxxxxxxx/.local/share/uv/python/cpython-3.10.16-linux-x86_64-gnu/lib/python3.10/asyncio/tasks.py:387]>
Exception ignored in: <coroutine object Queue.get at 0x7d585480ace0>
Traceback (most recent call last):
  File "/home/xxxxxxxxx/.local/share/uv/python/cpython-3.10.16-linux-x86_64-gnu/lib/python3.10/asyncio/queues.py", line 161, in get
    getter.cancel()  # Just in case getter is not done yet.
  File "/home/xxxxxxxxx/.local/share/uv/python/cpython-3.10.16-linux-x86_64-gnu/lib/python3.10/asyncio/base_events.py", line 753, in call_soon
    self._check_closed()
  File "/home/xxxxxxxxx/.local/share/uv/python/cpython-3.10.16-linux-x86_64-gnu/lib/python3.10/asyncio/base_events.py", line 515, in _check_closed
    raise RuntimeError('Event loop is closed')
RuntimeError: Event loop is closed

```

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)
This commit is contained in:
Yongteng Lei
2025-11-05 19:15:27 +08:00
committed by GitHub
parent 4e76220e25
commit 17ea5c1dee

View File

@ -61,7 +61,8 @@ class MCPToolCallSession(ToolCallSession):
for h, v in raw_headers.items():
nh = Template(h).safe_substitute(self._server_variables)
nv = Template(v).safe_substitute(self._server_variables)
headers[nh] = nv
if nh.strip() and nv.strip().strip("Bearer"):
headers[nh] = nv
if self._mcp_server.server_type == MCPServerType.SSE:
# SSE transport
@ -76,6 +77,9 @@ class MCPToolCallSession(ToolCallSession):
msg = f"Timeout initializing client_session for server {self._mcp_server.id}"
logging.error(msg)
await self._process_mcp_tasks(None, msg)
except asyncio.CancelledError:
logging.warning(f"SSE transport MCP session cancelled for server {self._mcp_server.id}")
return
except Exception:
msg = "Connection failed (possibly due to auth error). Please check authentication settings first"
await self._process_mcp_tasks(None, msg)
@ -93,6 +97,9 @@ class MCPToolCallSession(ToolCallSession):
msg = f"Timeout initializing client_session for server {self._mcp_server.id}"
logging.error(msg)
await self._process_mcp_tasks(None, msg)
except asyncio.CancelledError:
logging.warning(f"STREAMABLE_HTTP MCP session cancelled for server {self._mcp_server.id}")
return
except Exception as e:
logging.exception(e)
msg = "Connection failed (possibly due to auth error). Please check authentication settings first"
@ -107,6 +114,8 @@ class MCPToolCallSession(ToolCallSession):
mcp_task, arguments, result_queue = await asyncio.wait_for(self._queue.get(), timeout=1)
except asyncio.TimeoutError:
continue
except asyncio.CancelledError:
break
logging.debug(f"Got MCP task {mcp_task} arguments {arguments}")
@ -114,7 +123,10 @@ class MCPToolCallSession(ToolCallSession):
if not client_session or error_message:
r = ValueError(error_message)
await result_queue.put(r)
try:
await result_queue.put(r)
except asyncio.CancelledError:
break
continue
try:
@ -126,10 +138,18 @@ class MCPToolCallSession(ToolCallSession):
r = ValueError(f"Unknown MCP task {mcp_task}")
except Exception as e:
r = e
except asyncio.CancelledError:
break
await result_queue.put(r)
try:
await result_queue.put(r)
except asyncio.CancelledError:
break
async def _call_mcp_server(self, task_type: MCPTaskType, timeout: float | int = 8, **kwargs) -> Any:
if self._close:
raise ValueError("Session is closed")
results = asyncio.Queue()
await self._queue.put((task_type, kwargs, results))
@ -163,6 +183,9 @@ class MCPToolCallSession(ToolCallSession):
raise
def get_tools(self, timeout: float | int = 10) -> list[Tool]:
if self._close:
raise ValueError("Session is closed")
future = asyncio.run_coroutine_threadsafe(self._get_tools_from_mcp_server(timeout=timeout), self._event_loop)
try:
return future.result(timeout=timeout)
@ -176,6 +199,9 @@ class MCPToolCallSession(ToolCallSession):
@override
def tool_call(self, name: str, arguments: dict[str, Any], timeout: float | int = 10) -> str:
if self._close:
return "Error: Session is closed"
future = asyncio.run_coroutine_threadsafe(self._call_mcp_tool(name, arguments), self._event_loop)
try:
return future.result(timeout=timeout)
@ -191,8 +217,29 @@ class MCPToolCallSession(ToolCallSession):
return
self._close = True
self._event_loop.call_soon_threadsafe(self._event_loop.stop)
self._thread_pool.shutdown(wait=True)
while not self._queue.empty():
try:
_, _, result_queue = self._queue.get_nowait()
try:
await result_queue.put(asyncio.CancelledError("Session is closing"))
except Exception:
pass
except asyncio.QueueEmpty:
break
except Exception:
break
try:
self._event_loop.call_soon_threadsafe(self._event_loop.stop)
except Exception:
pass
try:
self._thread_pool.shutdown(wait=True)
except Exception:
pass
self.__class__._ALL_INSTANCES.discard(self)
def close_sync(self, timeout: float | int = 5) -> None:
@ -200,13 +247,16 @@ class MCPToolCallSession(ToolCallSession):
logging.warning(f"Event loop already stopped for {self._mcp_server.id}")
return
future = asyncio.run_coroutine_threadsafe(self.close(), self._event_loop)
try:
future.result(timeout=timeout)
except FuturesTimeoutError:
logging.error(f"Timeout while closing session for server {self._mcp_server.id} (timeout={timeout})")
future = asyncio.run_coroutine_threadsafe(self.close(), self._event_loop)
try:
future.result(timeout=timeout)
except FuturesTimeoutError:
logging.error(f"Timeout while closing session for server {self._mcp_server.id} (timeout={timeout})")
except Exception:
logging.exception(f"Unexpected error during close_sync for {self._mcp_server.id}")
except Exception:
logging.exception(f"Unexpected error during close_sync for {self._mcp_server.id}")
logging.exception(f"Exception while scheduling close for server {self._mcp_server.id}")
def close_multiple_mcp_toolcall_sessions(sessions: list[MCPToolCallSession]) -> None:
@ -215,16 +265,24 @@ def close_multiple_mcp_toolcall_sessions(sessions: list[MCPToolCallSession]) ->
async def _gather_and_stop() -> None:
try:
await asyncio.gather(*[s.close() for s in sessions if s is not None], return_exceptions=True)
except Exception:
logging.exception("Exception during MCP session cleanup")
finally:
loop.call_soon_threadsafe(loop.stop)
try:
loop.call_soon_threadsafe(loop.stop)
except Exception:
pass
loop = asyncio.new_event_loop()
thread = threading.Thread(target=loop.run_forever, daemon=True)
thread.start()
try:
loop = asyncio.new_event_loop()
thread = threading.Thread(target=loop.run_forever, daemon=True)
thread.start()
asyncio.run_coroutine_threadsafe(_gather_and_stop(), loop).result()
asyncio.run_coroutine_threadsafe(_gather_and_stop(), loop).result()
thread.join()
except Exception:
logging.exception("Exception during MCP session cleanup thread management")
thread.join()
logging.info(f"{len(sessions)} MCP sessions has been cleaned up. {len(list(MCPToolCallSession._ALL_INSTANCES))} in global context.")