全局唯一ID

全局唯一ID生成策略：

• UUID
• Redis自增
• snowflake算法
• 數(shù)據(jù)庫自增
  Redis自增ID策略：
• 每天一個key，方便統(tǒng)計訂單量
• ID構(gòu)造是 時間戳 + 計數(shù)器

@Component
public class RedisIdWorker {// 2024的第一時刻private static final long BEGIN_TIMESTAMP = 1704067200L;private static final int COUNT_BITS = 32;private final StringRedisTemplate stringRedisTemplate;public RedisIdWorker(StringRedisTemplate stringRedisTemplate) {this.stringRedisTemplate = stringRedisTemplate;}public long nextId(String keyPrefix){// 1. 獲取當(dāng)前時間戳LocalDateTime now = LocalDateTime.now();long nowSecond = now.toEpochSecond(ZoneOffset.UTC);long timeStamp = nowSecond - BEGIN_TIMESTAMP;// 2. 獲取序列號// 2.1 獲取當(dāng)天日期，精確到天String date = now.format(DateTimeFormatter.ofPattern("yyyy:MM:dd"));// 2.2 自增Long count = stringRedisTemplate.opsForValue().increment("icr:" + keyPrefix + ":" + date);// 3. 拼接成IDreturn timeStamp << COUNT_BITS | count;}}

超賣問題

在處理大量請求時，可能會出現(xiàn)超賣問題。
可以通過加鎖解決。

這里采用的是樂觀鎖。

一人一單

該任務(wù)需要每名用戶只能搶到一張優(yōu)惠券。
同時還要考慮到后端部署在多個服務(wù)器上可能會出現(xiàn)的異常，此時需要使用分布式鎖進(jìn)行解決。
這里的分布式鎖基于Redis實現(xiàn)。

基于Redis的分布式鎖實現(xiàn)思路

基于Redis的分布式鎖實現(xiàn)思路：

• 利用set nx ex獲取鎖，并設(shè)置過期時間，保存線程標(biāo)示
• 釋放鎖時先判斷線程標(biāo)示是否與自己一致，一致則刪除鎖

特性：

• 利用set nx滿足互斥性
• 利用set ex保證故障時鎖依然能釋放，避免死鎖，提高安全性
• 利用Redis集群保證高可用和高并發(fā)特性

使用Redis優(yōu)化秒殺

這里將庫存判斷與一人一單的校驗使用Redis完成。
具體流程為：

• 新增秒殺優(yōu)惠券的同時，將優(yōu)惠券信息保存到Redis中
• 基于Lua腳本，判斷秒殺庫存、一人一單，決定用戶是否搶購成功
• 如果搶購成功，將優(yōu)惠券id和用戶id封裝后存入阻塞隊列
• 開啟線程任務(wù)，不斷從阻塞隊列中獲取信息，實現(xiàn)異步下單功能
  這里的阻塞隊列是基于Stream的消息隊列
  STREAM類型消息隊列的XREADGROUP命令特點：
• 消息可回溯
• 可以多消費者爭搶消息，加快消費速度
• 可以阻塞讀取
• 沒有消息漏讀的風(fēng)險
• 有消息確認(rèn)機(jī)制，保證消息至少被消費一次
  
  新增秒殺優(yōu)惠券的同時，將優(yōu)惠券信息保存到Redis中

    @Override@Transactionalpublic void addSeckillVoucher(Voucher voucher) {// 保存優(yōu)惠券save(voucher);// 保存秒殺信息SeckillVoucher seckillVoucher = new SeckillVoucher();seckillVoucher.setVoucherId(voucher.getId());seckillVoucher.setStock(voucher.getStock());seckillVoucher.setBeginTime(voucher.getBeginTime());seckillVoucher.setEndTime(voucher.getEndTime());seckillVoucherService.save(seckillVoucher);// 保存優(yōu)惠券至RedisstringRedisTemplate.opsForValue().set(RedisConstants.SECKILL_STOCK_KEY +voucher.getId(), voucher.getStock().toString());}

基于Lua腳本，判斷秒殺庫存、一人一單，決定用戶是否搶購成功

-- 1.參數(shù)列表
local voucherId = ARGV[1]
local userId = ARGV[2]
local orderId = ARGV[3]
-- 2. 數(shù)據(jù)key
local stockKey = "seckill:stock:" .. voucherId
local orderKey = "seckill:order:" .. voucherId-- 3. 判斷庫存是否充足
if tonumber(redis.call('get', stockKey) )< 1 then-- 庫存不足return 1
end-- 4. 判斷用戶是否已經(jīng)搶購過
if redis.call('sismember', orderKey, userId) == 1 then-- 已經(jīng)搶購過return 2
end-- 5. 減庫存
redis.call('incrby', stockKey, -1)
-- 6. 記錄用戶搶購信息
redis.call('sadd', orderKey, userId)redis.call('xadd', "stream.orders", "*", "userId", userId,  "voucherId", voucherId,"id", orderId)
return 0

異步下單
LUA腳本

-- 1.參數(shù)列表
local voucherId = ARGV[1]
local userId = ARGV[2]
local orderId = ARGV[3]
-- 2. 數(shù)據(jù)key
local stockKey = "seckill:stock:" .. voucherId
local orderKey = "seckill:order:" .. voucherId-- 3. 判斷庫存是否充足
if tonumber(redis.call('get', stockKey) )< 1 then-- 庫存不足return 1
end-- 4. 判斷用戶是否已經(jīng)搶購過
if redis.call('sismember', orderKey, userId) == 1 then-- 已經(jīng)搶購過return 2
end-- 5. 減庫存
redis.call('incrby', stockKey, -1)
-- 6. 記錄用戶搶購信息
redis.call('sadd', orderKey, userId)redis.call('xadd', "stream.orders", "*", "userId", userId,  "voucherId", voucherId,"id", orderId)
return 0

從消息隊列取出，處理代碼

@Slf4j
@Service
public class VoucherOrderServiceImpl extends ServiceImpl<VoucherOrderMapper, VoucherOrder> implements IVoucherOrderService {@Resourceprivate ISeckillVoucherService seckillVoucherService;@Resourceprivate RedisIdWorker redisIdWorker;@Resourceprivate StringRedisTemplate stringRedisTemplate;@Resourceprivate RedissonClient redissonClient;private IVoucherOrderService proxy;// 靜態(tài)代碼塊加載Lua腳本private static final DefaultRedisScript<Long> SECKILL_SCRIPT;static {SECKILL_SCRIPT = new DefaultRedisScript<>();SECKILL_SCRIPT.setLocation(new ClassPathResource("seckill.lua"));SECKILL_SCRIPT.setResultType(Long.class);}//    private BlockingQueue<VoucherOrder> orderTasks = new ArrayBlockingQueue<>(1024*1024);// 線程池private static final ExecutorService SECKILL_ORDER_EXECUTOR = Executors.newSingleThreadExecutor();@PostConstruct// 完成類的construct即執(zhí)行下面的函數(shù)public void init() {// 交給線程池做SECKILL_ORDER_EXECUTOR.submit(new VoucherOrderHandler());}String queueName = "stream.orders";// 消費線程private class VoucherOrderHandler implements Runnable {@Overridepublic void run() {while (true) {try {// 4.1從消息隊列中取出訂單 xreadgroupnngroup g1 c1 count 1 block 200 streams streams.order >List<MapRecord<String, Object, Object>> list = stringRedisTemplate.opsForStream().read(Consumer.from("g1", "c1"),StreamReadOptions.empty().count(1).block(Duration.ofSeconds(2)),StreamOffset.create(queueName, ReadOffset.lastConsumed()));// 判斷是否成功// 沒有if (list == null || list.isEmpty()) {continue;}// 有// 4.2創(chuàng)建訂單,解析消息MapRecord<String, Object, Object> record = list.get(0);Map<Object, Object> value = record.getValue();VoucherOrder voucherOrder = BeanUtil.fillBeanWithMap(value, new VoucherOrder(), true);handleVoucherOrder(voucherOrder);// ACK確認(rèn)stringRedisTemplate.opsForStream().acknowledge(queueName, "g1", record.getId());} catch (Exception e) {log.error("訂單處理失敗", e);handlePendingList();}}}}private void handlePendingList() {while (true) {try {// 4.1從消息隊列中取出訂單 xreadgroupnngroup g1 c1 count 1 block 200 streams streams.order >List<MapRecord<String, Object, Object>> list = stringRedisTemplate.opsForStream().read(Consumer.from("g1", "c1"),StreamReadOptions.empty().count(1),StreamOffset.create(queueName, ReadOffset.from("0")));// 判斷是否成功// 沒有if (list == null || list.isEmpty()) {// pendingList 沒有消息break;}// 有// 4.2創(chuàng)建訂單,解析消息MapRecord<String, Object, Object> record = list.get(0);Map<Object, Object> value = record.getValue();VoucherOrder voucherOrder = BeanUtil.fillBeanWithMap(value, new VoucherOrder(), true);handleVoucherOrder(voucherOrder);// ACK確認(rèn)stringRedisTemplate.opsForStream().acknowledge(queueName, "g1", record.getId());} catch (Exception e) {log.error("訂單處理失敗", e);try {Thread.sleep(200);} catch (InterruptedException ex) {throw new RuntimeException(ex);}}}}private void handleVoucherOrder(VoucherOrder voucherOrder) {// 因為為子線程，userId只能從數(shù)據(jù)中取Long userId = voucherOrder.getUserId();RLock lock = redissonClient.getLock("order:" + userId);boolean isLock = lock.tryLock();if (!isLock) {log.error("重復(fù)搶購");return ;}try {proxy.createVoucherOrder(voucherOrder);return;} finally {lock.unlock();}}

創(chuàng)建訂單

    @Transactionalpublic Result createVoucherOrder(VoucherOrder voucherOrder) {// 4.一人一單Long userId = voucherOrder.getUserId();// 4.1 查詢訂單int count = query().eq("user_id", userId).eq("voucher_id", voucherOrder.getVoucherId()).count();if (count > 0) {return Result.fail("每人限購一張");}// 4.是// 4.1扣減庫存，基于樂觀鎖boolean success = seckillVoucherService.update().setSql("stock=stock-1").eq("voucher_id", voucherOrder.getVoucherId()).gt("stock", 0).update();if (!success) {return Result.fail("庫存不足");}//4.2創(chuàng)建訂單save(voucherOrder);//4.3返回訂單idreturn Result.ok(voucherOrder.getId());}

    @Overridepublic Result seckillVoucher(Long voucherId) {Long userId = UserHolder.getUser().getId();// 0. 生成訂單idlong orderId = redisIdWorker.nextId("order");// 1. 執(zhí)行l(wèi)ua腳本Long result = stringRedisTemplate.execute(SECKILL_SCRIPT,Collections.emptyList(),voucherId.toString(), userId.toString(), String.valueOf(orderId));// 2. 結(jié)果是否為0int r = result.intValue();// 2.1 不為0if (r!=0) {return Result.fail(r == 1 ? "庫存不足" : "不能重復(fù)搶購");}// 注解底層基于aop實現(xiàn)，需要獲得代理對象，進(jìn)行執(zhí)行proxy = (IVoucherOrderService)AopContext.currentProxy();// 3. 返回結(jié)果訂單idreturn Result.ok(orderId);}