Hmbown/CodeWhale · error

Fleet task does not exist

Error message

Fleet task {} does not exist

What it means

During restart_worker's coordinated restart callback, the manager searches the run's task_specs for the spec matching the task's task_id and fails when no spec matches. The ledger knows the task ran, but the run definition no longer carries a spec with that id — an inconsistency between durable task entries and the run's declared workload.

Solutions

1. Restore the original run definition so task_specs contains the referenced task id
2. Re-enqueue the task under the current run's spec ids instead of restarting the old one
3. Reconcile ledger task ids with run.task_specs before calling restart_worker
4. Check whether the run was re-created with new spec ids and restart by the new id

Example fix

// before
let spec = run.task_specs.iter().find(|s| s.id == task_id).ok_or(...)?;
// after
if !run.task_specs.iter().any(|s| s.id == task_id) {
    anyhow::bail!("task {task_id} was removed from the run; re-enqueue it");
}
let spec = run.task_specs.iter().find(|s| s.id == task_id).unwrap();

fn task_spec_exists(state: &FleetLedgerState, run_id: &str, task_id: &str) -> bool {
    state.runs.get(run_id)
        .map(|r| r.task_specs.iter().any(|s| s.id == task_id))
        .unwrap_or(false)
}

When it happens

Trigger: Calling restart_worker(worker_id) when the task's entry.task_id is not present in run.task_specs — e.g. the run manifest was rewritten, tasks were removed, or the wrong task id was recorded in the ledger.

Common situations: Editing a fleet run definition while tasks from an older definition are still in the ledger; copying task entries between runs; hand-editing run JSON.

Understand the failure class

Background: Record Not Found Errors: "not found", RecordNotFound, and "was not found" — what they mean and how to fix them — this error's family across 28 libraries.

Related errors

• attempt finalization requires a terminal worker event
• attempt finalization status must be terminal
• attempt receipt generation does not match its lease
• attempt receipt identity does not match its terminal event
• conditional progress append does not accept terminal worker…

AI-assisted analysis of Hmbown/CodeWhale@73e0f67d83 (2026-09-22). Data as JSON: /api/errors/37c104f4cd9ffea1. Report an issue: GitHub.

Appendix: source

Thrown at crates/tui/src/fleet/manager.rs:1119

        let restarted = self.ledger.restart_task_if_unchanged_with_callback(
            &task.entry.run_id,
            &task.entry.task_id,
            worker_id,
            task.status,
            task.entry.attempts,
            latest_seq,
            heartbeat_at,
            &now,
            None,
            task.entry.attempts,
            || {
                if let Some(guard) = coordination_guard.as_mut() {
                    let task_spec = run
                        .task_specs
                        .iter()
                        .find(|spec| spec.id == task.entry.task_id)
                        .ok_or_else(|| {
                            anyhow!("Fleet task {} does not exist", task.entry.task_id)
                        })?;
                    self.prepare_registered_restart_generation(
                        guard, &state, task, task_spec, worker_id,
                    )?;
                }
                Ok(())
            },
        )?;
        if !restarted {
            bail!("worker {worker_id} task changed before it could be restarted");
        }
        self.ledger
            .update_run_status(&task.entry.run_id, FleetRunStatus::Running, &timestamp())?;
        Ok(FleetRestartReport {
            run_id: task.entry.run_id.clone(),
            max_workers,
            inspection: self.inspect_worker(worker_id)?,
        })

View on GitHub (pinned to 73e0f67d83)