What causes "Python on_stop failed: {e}" in nautechsystems/nautilus_trader?

Read the chained `{e}` traceback and fix the Python on_stop implementation.

Python on_stop failed: {e}

Error message

Python on_stop failed: {e}

What it means

The DataActor on_stop path first runs Rust ExecutionAlgorithm::on_stop, then dispatches Python `on_stop`; a Python exception is re-wrapped with this message. It means the Python component failed during orderly shutdown.

Source

Thrown at crates/trading/src/python/algorithm.rs:553

            .dispatch_position_event("on_position_closed", PositionEvent::PositionClosed(event));
    }

    fn on_position_event(&mut self, event: PositionEvent) {
        let _ = self.dispatch_position_event("on_position_event", event);
    }
}

impl DataActor for PyExecutionAlgorithm {
    fn on_start(&mut self) -> anyhow::Result<()> {
        ExecutionAlgorithm::on_start(self)?;
        self.dispatch_no_args("on_start")
            .map_err(|e| anyhow::anyhow!("Python on_start failed: {e}"))
    }

    fn on_stop(&mut self) -> anyhow::Result<()> {
        ExecutionAlgorithm::on_stop(self)?;
        self.dispatch_no_args("on_stop")
            .map_err(|e| anyhow::anyhow!("Python on_stop failed: {e}"))
    }

    fn on_resume(&mut self) -> anyhow::Result<()> {
        ExecutionAlgorithm::on_resume(self)?;
        self.dispatch_no_args("on_resume")
            .map_err(|e| anyhow::anyhow!("Python on_resume failed: {e}"))
    }

    fn on_reset(&mut self) -> anyhow::Result<()> {
        ExecutionAlgorithm::on_reset(self)?;
        self.dispatch_no_args("on_reset")
            .map_err(|e| anyhow::anyhow!("Python on_reset failed: {e}"))
    }

    fn on_dispose(&mut self) -> anyhow::Result<()> {
        self.dispatch_no_args("on_dispose")
            .map_err(|e| anyhow::anyhow!("Python on_dispose failed: {e}"))
    }

View on GitHub (pinned to 18893faf8b)

Solutions

Read the chained `{e}` traceback and fix the Python on_stop implementation.
Make on_stop idempotent and defensive (guard already-closed/already-stopped resources).
Ensure state required by on_stop is not destroyed before stop is invoked.
Verify the Python class defines a zero-argument on_stop.

Example fix

// before
def on_stop(self):
    self.conn.close()  # raises if already closed

// after
def on_stop(self):
    if self.conn is not None and not self.conn.closed:
        self.conn.close()

Defensive patterns

Strategy: try-catch

Validate before calling

# Python: make stop idempotent before shutdown
if getattr(self, "stopped", False): return

Try / catch

if let Err(e) = actor.on_stop() {
    log::error!("python on_stop failed: {e:#}"); // still proceed with teardown
}

Prevention

Write on_stop defensively: wrap each cleanup step so one failure doesn't abort the rest.
Track a stopped/disposed flag to make shutdown hooks idempotent.
Keep shutdown-time state (handlers, connections) alive until on_stop completes.

When it happens

Trigger: Calling on_stop when the Python on_stop callback raises — e.g. errors while flushing/closing resources, or a bad no-arg dispatch.

Common situations: Closing an already-closed connection in on_stop; exception during final state persistence; callbacks that reference handlers cleared during shutdown.

Related errors

AI-assisted analysis of nautechsystems/nautilus_trader@18893faf8b (2026-09-08). Data as JSON: /api/errors/7b73a4f3228aa9fd. Report an issue: GitHub.