RocketRide server shutdown/restart can raise a Uvicorn signal-handler restoration error:
Observed Stack
ai/eaas.py
main()
asyncio.run(run())
ai/eaas.py
await server.serve()
ai/web/server.py
await self.server.serve()
uvicorn/server.py
with self.capture_signals()
uvicorn/server.py
signal.signal(sig, handler)
python3.12/signal.py
TypeError
TypeError: signal handler must be signal.SIG_IGN, signal.SIG_DFL, or a callable object```
The server can restart successfully afterward, but during the restart/shutdown path is noisy with the signal exceptions.
### Steps to Reproduce
This appears during RocketRide normal server shutdown/restart, not during normal pipeline node execution.
### Expected vs Actual Behavior
Expected Behavior
Server shutdown/restart should complete cleanly without raising a signal restoration exception.
Actual Behavior
During shutdown/restart, Uvicorn attempts to restore a saved signal handler value that Python rejects because it is not SIG_IGN, SIG_DFL, or callable.
### Severity
P3 - Low (minor inconvenience)
### Platform
- [ ] Windows
- [ ] macOS
- [ ] Linux
### Version / Environment
_No response_
### Screenshots / Logs
_No response_
Description
Description
RocketRide server shutdown/restart can raise a Uvicorn signal-handler restoration error: