Files
hi-server/queue/logic/order/stuckOrderRecoveryLogic.go
shanshanzhong147 30221232c9 fix: 修复订单状态机 claim 机制的三个 Bug 并增加 stuck 订单恢复
- 将 OrderStatusClaimed 从 4 改为 6,消除与 OrderStatusFailed 的值冲突
- finalizeCouponAndOrder 改用直接 DB 更新(WHERE status=6→SET status=5),
  绕过 UpdateOrderStatus 的 status<target 守卫,同时用 model.Update 刷新缓存
- releaseClaim 返回 error,调用处检查并记录日志;releaseClaim 失败由 stuck
  recovery 定时任务兜底
- claimAndGetOrder 对 status=claimed 返回可重试错误而非静默跳过;
  ProcessTask 区分 "stuck in claimed" 与 "非 paid 跳过" 两种场景
- 新增 StuckOrderRecoveryLogic:每 10 分钟扫描超时 claimed 订单,
  重置 status=paid 并重新入队 ForthwithActivateOrder,确保不依赖
  asynq 原始重试(可能已超 maxRetry)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: multica-agent <github@multica.ai>
2026-05-25 02:18:19 -07:00

105 lines
3.3 KiB
Go

package orderLogic
import (
"context"
"encoding/json"
"time"
"github.com/hibiken/asynq"
"github.com/perfect-panel/server/internal/model/order"
"github.com/perfect-panel/server/internal/svc"
"github.com/perfect-panel/server/pkg/logger"
queueTypes "github.com/perfect-panel/server/queue/types"
)
// StuckOrderRecoveryLogic scans orders stuck in claimed status and re-queues them for processing.
type StuckOrderRecoveryLogic struct {
svc *svc.ServiceContext
}
func NewStuckOrderRecoveryLogic(svc *svc.ServiceContext) *StuckOrderRecoveryLogic {
return &StuckOrderRecoveryLogic{svc: svc}
}
// ProcessTask scans for orders stuck in claimed status for over 10 minutes,
// resets them to paid, and re-enqueues the activate task so they are retried
// independently of asynq's original retry counter (which may be exhausted).
func (l *StuckOrderRecoveryLogic) ProcessTask(ctx context.Context, _ *asynq.Task) error {
cutoff := time.Now().Add(-10 * time.Minute)
var stuckOrders []order.Order
if err := l.svc.DB.WithContext(ctx).
Model(&order.Order{}).
Where("status = ? AND updated_at < ?", OrderStatusClaimed, cutoff).
Find(&stuckOrders).Error; err != nil {
logger.WithContext(ctx).Error("[StuckOrderRecovery] Failed to query stuck orders",
logger.Field("error", err.Error()),
)
return err
}
if len(stuckOrders) == 0 {
return nil
}
orderNos := make([]string, 0, len(stuckOrders))
for i := range stuckOrders {
orderNos = append(orderNos, stuckOrders[i].OrderNo)
}
logger.WithContext(ctx).Error("[StuckOrderRecovery] Found stuck claimed orders, recovering",
logger.Field("count", len(stuckOrders)),
logger.Field("order_nos", orderNos),
)
for i := range stuckOrders {
o := &stuckOrders[i]
result := l.svc.DB.WithContext(ctx).
Model(&order.Order{}).
Where("order_no = ? AND status = ?", o.OrderNo, OrderStatusClaimed).
Update("status", OrderStatusPaid)
if result.Error != nil {
logger.WithContext(ctx).Error("[StuckOrderRecovery] Failed to reset order status",
logger.Field("order_no", o.OrderNo),
logger.Field("error", result.Error.Error()),
)
continue
}
if result.RowsAffected == 0 {
// Another process already handled this order
continue
}
// Invalidate order cache
o.Status = OrderStatusPaid
if err := l.svc.OrderModel.Update(ctx, o); err != nil {
logger.WithContext(ctx).Error("[StuckOrderRecovery] Failed to update order cache",
logger.Field("order_no", o.OrderNo),
logger.Field("error", err.Error()),
)
}
// Re-enqueue activate task so the order gets processed regardless of asynq retry state
payload, err := json.Marshal(queueTypes.ForthwithActivateOrderPayload{OrderNo: o.OrderNo})
if err != nil {
logger.WithContext(ctx).Error("[StuckOrderRecovery] Failed to marshal task payload",
logger.Field("order_no", o.OrderNo),
logger.Field("error", err.Error()),
)
continue
}
if _, err = l.svc.Queue.EnqueueContext(ctx, asynq.NewTask(queueTypes.ForthwithActivateOrder, payload)); err != nil {
logger.WithContext(ctx).Error("[StuckOrderRecovery] Failed to re-enqueue activate task",
logger.Field("order_no", o.OrderNo),
logger.Field("error", err.Error()),
)
} else {
logger.WithContext(ctx).Info("[StuckOrderRecovery] Re-enqueued activate task",
logger.Field("order_no", o.OrderNo),
)
}
}
return nil
}