Hmbown/CodeWhale · error

Failed to persist turn: {persistence_error}; cleanup also fa

Error message

Failed to persist turn: {persistence_error}; cleanup also failed: {cleanup_error}

What it means

Raised in submit_turn after the durable-store closure (save_item/save_turn/save_thread) returned Err for a newly constructed user turn. It is a compound sentinel: the primary persistence_error says why the store write failed, and cleanup_error is appended only because the compensating cleanup_unaccepted_turn_records call that tries to roll back the half-written item/turn also failed, leaving the store possibly inconsistent.

Source

Thrown at crates/tui/src/runtime_threads.rs:6147

            state.route_model.clone_from(&model);

            let persistence_result = (|| -> Result<()> {
                self.store.save_item(&user_item)?;
                self.store.save_turn(&turn)?;
                current_thread.latest_turn_id = Some(turn_id.clone());
                current_thread.updated_at = now;
                self.store.save_thread(&current_thread)
            })();
            if let Err(persistence_error) = persistence_result {
                let cleanup_error = self
                    .cleanup_unaccepted_turn_records(&turn_id, Some(&user_item_id))
                    .err();
                state.active_turn = None;
                state.route_identity = previous_active_route.0;
                state.route_model = previous_active_route.1;
                return match cleanup_error {
                    None => Err(anyhow!("Failed to persist turn: {persistence_error}")),
                    Some(cleanup_error) => Err(anyhow!(
                        "Failed to persist turn: {persistence_error}; cleanup also failed: {cleanup_error}"
                    )),
                };
            }

            self.register_runtime_usage_sink(&turn_id);
            // Sending through an owned permit cannot await or fail. From this
            // point the engine owns the operation and the spawned task owns
            // lifecycle events, monitoring, and terminal cleanup even if the
            // HTTP/client future is dropped.
            engine.publish_turn_authority(
                mode,
                allow_shell,
                trust_mode,
                auto_approve,
                policy.permission,
                configured_sandbox_mode,
            );

View on GitHub (pinned to 0c42157ee5)

Solutions

1. Inspect persistence_error first (disk full, store lock, schema corruption) and fix that root cause; the cleanup error is secondary noise from the rollback attempt
2. If cleanup also failed, check the store for orphaned records with the reported turn id / user_item id and delete them manually before retrying the turn
3. Verify the store directory is writable and not concurrently locked by another Codewhale process, then resubmit the turn

When it happens

Trigger: Thrown at crates/tui/src/runtime_threads.rs:6147 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/97f6750a04716cbb. Report an issue: GitHub.