跟接上文:
并发情况下的库存扣减

一 库存扣减

public DefaultTreeFactory.TreeActionEntity logic(String userId, Long strategyId, Integer awardId, String ruleValue, Date endDateTime) {
        log.info("规则过滤-库存扣减 userId:{} strategyId:{} awardId:{}", userId, strategyId, awardId);
        // 扣减库存
        Boolean status = strategyDispatch.subtractionAwardStock(strategyId, awardId, endDateTime);
        // true；库存扣减成功，TAKE_OVER 规则节点接管，返回奖品ID，奖品规则配置
        if (status) {
            log.info("规则过滤-库存扣减-成功 userId:{} strategyId:{} awardId:{}", userId, strategyId, awardId);

            // 写入延迟队列，延迟消费更新数据库记录。【在trigger的job；UpdateAwardStockJob 下消费队列，更新数据库记录】
            strategyRepository.awardStockConsumeSendQueue(StrategyAwardStockKeyVO.builder()
                    .strategyId(strategyId)
                    .awardId(awardId)
                    .build());

            // 注意；根据数据库表中配置走不同的节点。目前数据库配置 ALLOW 是走到下一个节点。
            return DefaultTreeFactory.TreeActionEntity.builder()
                    .ruleLogicCheckType(RuleLogicCheckTypeVO.TAKE_OVER)
                    .strategyAwardVO(DefaultTreeFactory.StrategyAwardVO.builder()
                            .awardId(awardId)
                            .awardRuleValue(ruleValue)
                            .build())
                    .build();
        }

        // 如果库存不足，则直接返回放行
        log.warn("规则过滤-库存扣减-告警，库存不足。userId:{} strategyId:{} awardId:{}", userId, strategyId, awardId);
        return DefaultTreeFactory.TreeActionEntity.builder()
                .ruleLogicCheckType(RuleLogicCheckTypeVO.ALLOW)
                .build();
    }

public void awardStockConsumeSendQueue(StrategyAwardStockKeyVO strategyAwardStockKeyVO) {
        String cacheKey = Constants.RedisKey.STRATEGY_AWARD_COUNT_QUERY_KEY;
        RBlockingQueue<StrategyAwardStockKeyVO> blockingQueue = redisService.getBlockingQueue(cacheKey);
        RDelayedQueue<StrategyAwardStockKeyVO> delayedQueue = redisService.getDelayedQueue(blockingQueue);
        delayedQueue.offer(strategyAwardStockKeyVO, 3, TimeUnit.SECONDS);
    }

二 定时任务扫描扣减数据库库存

@Slf4j
@Component()
public class UpdateAwardStockJob {

    @Resource
    private IRaffleStock raffleStock;

//    @Scheduled(cron = "0/5 * * * * ?")
    public void exec() {
        try {
            StrategyAwardStockKeyVO strategyAwardStockKeyVO = raffleStock.takeQueueValue();
            if (null == strategyAwardStockKeyVO) return;
            log.info("定时任务，更新奖品消耗库存 strategyId:{} awardId:{}", strategyAwardStockKeyVO.getStrategyId(), strategyAwardStockKeyVO.getAwardId());
            raffleStock.updateStrategyAwardStock(strategyAwardStockKeyVO.getStrategyId(), strategyAwardStockKeyVO.getAwardId());
        } catch (Exception e) {
            log.error("定时任务，更新奖品消耗库存失败", e);
        }
    }

}

  public StrategyAwardStockKeyVO takeQueueValue() throws InterruptedException {
        String cacheKey = Constants.RedisKey.STRATEGY_AWARD_COUNT_QUERY_KEY;
        RBlockingQueue<StrategyAwardStockKeyVO> destinationQueue = redisService.getBlockingQueue(cacheKey);
        return destinationQueue.poll();
    }

<update id="updateStrategyAwardStock" parameterType="cn.bugstack.infrastructure.persistent.po.StrategyAward">
        update strategy_award
        set award_count_surplus = award_count_surplus - 1
        where strategy_id = #{strategyId} and award_id = #{awardId} and award_count_surplus > 0
 </update>

三 redis库存扣减到0,直接清空队列并修改数据库库存为0

@Autowired
    private RabbitTemplate rabbitTemplate;

    public void publish(String topic, BaseEvent.EventMessage<?> eventMessage) {
        try {
            String messageJson = JSON.toJSONString(eventMessage);
            rabbitTemplate.convertAndSend(topic, messageJson);
            log.info("发送MQ消息 topic:{} message:{}", topic, messageJson);
        } catch (Exception e) {
            log.error("发送MQ消息失败 topic:{} message:{}", topic, JSON.toJSONString(eventMessage), e);
            throw e;
        }
    }

@Slf4j
@Component
public class ActivitySkuStockZeroCustomer {

    @Value("${spring.rabbitmq.topic.activity_sku_stock_zero}")
    private String topic;

    @Resource
    private IRaffleActivitySkuStockService skuStock;

    @RabbitListener(queuesToDeclare = @Queue(value = "${spring.rabbitmq.topic.activity_sku_stock_zero}"))
    public void listener(String message) {
        try {
            log.info("监听活动sku库存消耗为0消息 topic: {} message: {}", topic, message);
            // 转换对象
            BaseEvent.EventMessage<Long> eventMessage = JSON.parseObject(message, new TypeReference<BaseEvent.EventMessage<Long>>() {
            }.getType());
            Long sku = eventMessage.getData();
            // 更新库存
            skuStock.clearActivitySkuStock(sku);
            // 清空队列 「此时就不需要延迟更新数据库记录了」todo 清空时，需要设定sku标识，不能全部清空。
            skuStock.clearQueueValue();
        } catch (Exception e) {
            log.error("监听活动sku库存消耗为0消息，消费失败 topic: {} message: {}", topic, message);
            throw e;
        }
    }

}

<update id="clearActivitySkuStock" parameterType="java.lang.Long">
        update raffle_activity_sku
        set stock_count_surplus = 0, update_time = now()
        where sku = #{sku}
    </update>

@Override
    public void clearQueueValue() {
        String cacheKey = Constants.RedisKey.ACTIVITY_SKU_COUNT_QUERY_KEY;
        RBlockingQueue<ActivitySkuStockKeyVO> destinationQueue = redisService.getBlockingQueue(cacheKey);
        destinationQueue.clear();
    }