vxcontrol/pentagi · error

failed to set subtask %d status to finished: %w

Error message

failed to set subtask %d status to finished: %w

What it means

Wraps the error returned by subtaskWorker.SetStatus when the worker tried to transition a subtask to 'finished' after the agent chain reported PerformResultDone. The database status update failed (likely context cancellation, DB connectivity loss, or a status conflict), so the subtask is left in a non-finished state even though the work succeeded. handleInterrupting is invoked first, so context.Canceled/DeadlineExceeded resets the subtask to Waiting for a later retry.

Source

Thrown at backend/pkg/controller/subtask.go:341

		}
		errChainConsistency := stw.subtaskCtx.Provider.EnsureChainConsistency(ctx, msgChainID)
		if errChainConsistency != nil {
			err = errors.Join(err, errChainConsistency)
		}
		_ = stw.SetStatus(ctx, database.SubtaskStatusWaiting)
		return fmt.Errorf("failed to perform agent chain for subtask %d: %w", subtaskID, err)
	}

	switch performResult {
	case providers.PerformResultWaiting:
		if err := stw.SetStatus(ctx, database.SubtaskStatusWaiting); err != nil {
			stw.handleInterrupting(err)
			return err
		}
	case providers.PerformResultDone:
		if err := stw.SetStatus(ctx, database.SubtaskStatusFinished); err != nil {
			stw.handleInterrupting(err)
			return fmt.Errorf("failed to set subtask %d status to finished: %w", subtaskID, err)
		}
	case providers.PerformResultError:
		if err := stw.SetStatus(ctx, database.SubtaskStatusFailed); err != nil {
			stw.handleInterrupting(err)
			return fmt.Errorf("failed to set subtask %d status to failed: %w", subtaskID, err)
		}
	default:
		return fmt.Errorf("unknown perform result: %d", performResult)
	}

	return nil
}

// handleInterrupting sets this subtask (and task/flow via SetStatus back-propagation)
// to Waiting when err is context.Canceled or context.DeadlineExceeded. Use after the subtask
// was advanced past Waiting (e.g. Running) but the run aborts before PerformAgentChain's
// normal error handler, or when a late SetStatus fails with a context interruption.
func (stw *subtaskWorker) handleInterrupting(err error) {

View on GitHub (pinned to ea665308ba)

Solutions

1. Check DB connectivity/pool health (pgx logs) and retry the flow; if the error is context.Canceled the worker already reset the subtask to Waiting, so re-run it.
2. Ensure the context passed to Run is not canceled before completion — don't abort the flow goroutine while subtasks are finishing.
3. Verify no concurrent code path updates the same subtask status (idempotency: finished→finished transitions or unique constraints failing).
4. Check the wrapped error (%w) in logs for the root cause (e.g. SQLSTATE) and address that specifically.

Example fix

// before: run with the flow's context that gets canceled on timeout
err := stw.Run(flowCtx)
// after: tolerate interruptions — worker resets to Waiting; re-dispatch
if err != nil {
    var subtaskErr *controller.SubtaskError
    if errors.Is(err, context.Canceled) {
        // subtask was reset to Waiting by handleInterrupting; safe to retry
        stc.TryRequeueSubtask(subtaskID)
    }
}

if stw.IsCompleted() { return nil } // never run a terminal subtask
if err := db.PingContext(ctx); err != nil { return fmt.Errorf("db unavailable: %w", err) }

func isInterrupting(err error) bool {
    return errors.Is(err, context.Canceled) || errors.Is(err, context.DeadlineExceeded)
}

if err := stw.Run(ctx); err != nil {
    if isInterrupting(err) {
        // worker reset subtask to Waiting; safe to requeue
        requeue(subtaskID)
        return nil
    }
    return fmt.Errorf("run failed: %w", err)
}

When it happens

Trigger: subtaskWorker.Run() gets providers.PerformResultDone from Provider.PerformAgentChain and the subsequent SetStatus(ctx, SubtaskStatusFinished) returns an error — e.g. the run context was canceled mid-update, the PostgreSQL connection dropped, or the row was concurrently modified.

Common situations: Flow cancellation or shutdown racing with task completion; DB pod restart/connection pool exhaustion in Kubernetes; pgx 'conn busy' or serialization failures under heavy concurrent flow updates; the parent task being deleted while its last subtask finishes.

Related errors

• failed to set subtask %d status to failed: %w
• failed to update subtask %d status to created: %w
• failed to get subtask primary msg chains for subtask %d: %w
• failed to set subtask %d status: %w
• failed to get subtasks for task %d: %w

AI-assisted analysis of vxcontrol/pentagi@ea665308ba (2026-09-01). Data as JSON: /api/errors/fa14e352922adefb. Report an issue: GitHub.