Stdio transport requires multiple Ctrl-C to terminate

[jlowin]

When running a FastMCP server with stdio transport directly via python server.py, terminating with Ctrl-C requires multiple presses and produces messy tracebacks.

Reproduction

from fastmcp import FastMCP

mcp = FastMCP("Test Server")

@mcp.tool
def hello(name: str) -> str:
    return f"Hello, {name}!"

if __name__ == "__main__":
    mcp.run()

Run with python server.py, then press Ctrl-C.

Expected behavior

Single Ctrl-C cleanly terminates the server.

Actual behavior

Requires 2-3 Ctrl-C presses and produces cascading KeyboardInterrupt tracebacks:

^C^CTraceback (most recent call last):
  File "..._asyncio.py", line 2365, in run
    return runner.run(wrapper())
  ...
  File ".../asyncio/runners.py", line 157, in _on_sigint
    raise KeyboardInterrupt()
KeyboardInterrupt

During handling of the above exception, another exception occurred:
...

Root cause

The stdio transport blocks on selector.select() waiting for stdin. When SIGINT arrives:

1. The blocking call is interrupted
2. Async cleanup begins but gets interrupted by subsequent Ctrl-C
3. Threading shutdown also gets interrupted

Notes

• HTTP transport handles Ctrl-C cleanly (uvicorn has proper signal handling)
• Stdio is designed to be controlled by a parent MCP client process, so interactive use is an edge case
• However, python server.py is a common way to test servers during development

[marvin-context-protocol]

Found 1 possible duplicate issue:

1. FastMCP process faills to terminate on KeyboardInterrupt #2788: FastMCP process faills to terminate on KeyboardInterrupt

This issue will be automatically closed as a duplicate in 3 days.

• If your issue is a duplicate, please close it and 👍 the existing issue instead
• To prevent auto-closure, add a comment or 👎 this comment