Hmbown/CodeWhale · error

scheduling Fleet run {} after draining owned workers

Error message

scheduling Fleet run {} after draining owned workers

What it means

This is an informational log (not a panic) in the Fleet manager loop: after observing that owned workers are draining, the manager schedules the run while explicitly excluding those worker ids (schedule_run_excluding) so no new work is leased onto a logical worker whose executor handle has not yet been observed and forgotten. It records the deliberate drain-then-schedule decision when a terminal ledger update races the foreground host process.

Source

Thrown at crates/tui/src/fleet/manager.rs:784

                            manager_lock_path.display()
                        )
                    });
                }
            }
        };
        loop {
            // A terminal ledger update can race the foreground host process.
            // Do not lease new work onto a logical worker until its executor
            // handle has been observed and forgotten below.
            let unavailable_workers = executor.worker_ids().into_iter().collect();
            let scheduling_error = self
                .schedule_run_excluding(run_id, max_workers, &unavailable_workers)
                .err();
            self.drive_executor_tick(run_id, executor, codewhale_binary, model)?;
            self.refresh_run_status(run_id)?;
            if let Some(error) = scheduling_error {
                if executor.worker_ids().is_empty() {
                    return Err(error).with_context(|| {
                        format!(
                            "scheduling Fleet run {} after draining owned workers",
                            run_id.0
                        )
                    });
                }
                tracing::warn!(
                    run_id = %run_id.0,
                    error = %error,
                    "Fleet scheduling paused while already-leased workers continue"
                );
            }
            // A separate `fleet interrupt` process can make the ledger
            // terminal while this manager still owns a live host child. Keep
            // driving until the executor has observed that cancellation and
            // stopped every tracked process.
            if !self.run_has_open_work(run_id)? && executor.worker_ids().is_empty() {
                return self.run_status(run_id);

View on GitHub (pinned to 0c42157ee5)

Solutions

1. No action needed on the log itself; it confirms scheduling continued on the remaining workers
2. If scheduling repeatedly happens after draining, investigate why workers keep terminating
3. Ensure executor.worker_ids() reflects forgotten handles so the exclusion list stays accurate

When it happens

Trigger: Thrown at crates/tui/src/fleet/manager.rs:784 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of Hmbown/CodeWhale@0c42157ee5 (2026-08-20). Data as JSON: /api/errors/ad7de4a921eacf6c. Report an issue: GitHub.