feat: add 50 backend interview questions and answers
Generated comprehensive interview preparation materials covering: - Distributed systems (transactions, locks, ID generation, consistency) - Database (indexing, sharding, replication, transactions) - Caching (Redis, cache problems, distributed lock) - Message queues (RocketMQ, Kafka) - Concurrency (ThreadLocal, ConcurrentHashMap, thread pools) - JVM (GC, memory, tuning) - System design (seckill, short URL, IM, feed, LBS) - Algorithms (B+ tree, LRU, Red-Black tree, Skip list, Timing wheel) - Network (TCP/IP, HTTP/HTTPS) - Security (encryption, SQL injection, XSS) - Performance tuning - Design patterns - Microservices (Spring Boot, Gateway, Service Mesh) - Container orchestration (Kubernetes, Docker) - CI/CD, observability Each file includes: - Detailed questions - Comprehensive answers - Code examples - Real project experience - Alibaba P7 level requirements Generated with [Claude Code](https://claude.com/claude-code) via [Happy](https://happy.engineering) Co-Authored-By: Claude <noreply@anthropic.com> Co-Authored-By: Happy <yesreply@happy.engineering>
This commit is contained in:
yasinshaw
456
questions/distributed-lock.md
Normal file
456
questions/distributed-lock.md
Normal file
@@ -0,0 +1,456 @@
|
||||
# 分布式锁
|
||||
|
||||
## 问题
|
||||
|
||||
1. 什么是分布式锁？为什么需要分布式锁？
|
||||
2. 如何用 Redis 实现分布式锁？
|
||||
3. Redis 分布式锁有哪些坑？如何解决？
|
||||
4. Zookeeper 如何实现分布式锁？
|
||||
5. Redis 分布式锁和 Zookeeper 分布式锁的区别？
|
||||
6. Redisson 是如何实现分布式锁的？
|
||||
7. 在实际项目中如何使用分布式锁？
|
||||
|
||||
---
|
||||
|
||||
## 标准答案
|
||||
|
||||
### 1. 为什么需要分布式锁
|
||||
|
||||
#### **场景**
|
||||
|
||||
**单机锁（synchronized、ReentrantLock）**：
|
||||
```java
|
||||
@Service
|
||||
public class OrderService {
|
||||
private synchronized void createOrder(Order order) {
|
||||
// 单机环境下有效
|
||||
// 分布式环境下无效（不同 JVM）
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
**问题**：
|
||||
- 单机锁只对单个 JVM 有效
|
||||
- 分布式环境下，多个实例的锁互不排斥
|
||||
- 需要跨 JVM、跨机器的锁机制
|
||||
|
||||
---
|
||||
|
||||
### 2. Redis 分布式锁
|
||||
|
||||
#### **基本实现**
|
||||
|
||||
```java
|
||||
public class RedisDistributedLock {
|
||||
|
||||
private final StringRedisTemplate redisTemplate;
|
||||
|
||||
// 加锁
|
||||
public boolean lock(String key, String value, long expireTime) {
|
||||
// SET key value NX PX expireTime
|
||||
Boolean success = redisTemplate.opsForValue()
|
||||
.setIfAbsent(key, value, expireTime, TimeUnit.MILLISECONDS);
|
||||
|
||||
return success != null && success;
|
||||
}
|
||||
|
||||
// 释放锁
|
||||
public boolean unlock(String key, String value) {
|
||||
// Lua 脚本：保证原子性
|
||||
String luaScript =
|
||||
"if redis.call('get', KEYS[1]) == ARGV[1] then " +
|
||||
" return redis.call('del', KEYS[1]) " +
|
||||
"else " +
|
||||
" return 0 " +
|
||||
"end";
|
||||
|
||||
DefaultRedisScript<Long> script = new DefaultRedisScript<>(luaScript, Long.class);
|
||||
Long result = redisTemplate.execute(script, Collections.singletonList(key), value);
|
||||
|
||||
return result != null && result == 1;
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
#### **使用示例**
|
||||
|
||||
```java
|
||||
@Service
|
||||
public class OrderService {
|
||||
|
||||
@Autowired
|
||||
private RedisDistributedLock lock;
|
||||
|
||||
public void createOrder(Order order) {
|
||||
String lockKey = "order:" + order.getProductId();
|
||||
String lockValue = UUID.randomUUID().toString();
|
||||
|
||||
try {
|
||||
// 加锁（过期时间 30 秒）
|
||||
boolean locked = lock.lock(lockKey, lockValue, 30000);
|
||||
|
||||
if (!locked) {
|
||||
throw new BusinessException("系统繁忙，请稍后再试");
|
||||
}
|
||||
|
||||
// 执行业务逻辑
|
||||
// 1. 查询库存
|
||||
// 2. 扣减库存
|
||||
// 3. 创建订单
|
||||
|
||||
} finally {
|
||||
// 释放锁
|
||||
lock.unlock(lockKey, lockValue);
|
||||
}
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 3. Redis 分布式锁的坑
|
||||
|
||||
#### **坑 1：锁超时问题**
|
||||
|
||||
**场景**：
|
||||
```
|
||||
线程 A 获取锁（设置 30 秒过期）
|
||||
线程 A 执行业务（耗时 50 秒）
|
||||
30 秒后，锁自动过期
|
||||
线程 B 获取锁
|
||||
线程 A 执行完毕，释放锁（把线程 B 的锁释放了！）
|
||||
```
|
||||
|
||||
**解决方案：看门狗（Watchdog）**
|
||||
|
||||
**原理**：
|
||||
- 启动后台线程
|
||||
- 定期检查锁是否还存在
|
||||
- 如果存在，续期（延长过期时间）
|
||||
|
||||
**Redisson 实现**：
|
||||
```java
|
||||
RLock lock = redisson.getLock("myLock");
|
||||
lock.lock(); // 自动续期（默认 30 秒）
|
||||
|
||||
try {
|
||||
// 业务逻辑
|
||||
} finally {
|
||||
lock.unlock();
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
#### **坑 2：主从切换导致锁丢失**
|
||||
|
||||
**场景**：
|
||||
```
|
||||
线程 A 在 Master 获取锁
|
||||
锁未同步到 Slave
|
||||
Master 宕机，Slave 升级为 Master
|
||||
线程 B 在新 Master 上获取锁（冲突！）
|
||||
```
|
||||
|
||||
**解决方案：Redlock（Redis 分布式锁算法）**
|
||||
|
||||
**原理**：
|
||||
- 向多个独立的 Redis 节点请求加锁
|
||||
- 大多数节点加锁成功才算成功
|
||||
|
||||
```java
|
||||
// Redisson Redlock
|
||||
RLock lock1 = redisson1.getLock("myLock");
|
||||
RLock lock2 = redisson2.getLock("myLock");
|
||||
RLock lock3 = redisson3.getLock("myLock");
|
||||
|
||||
RedissonRedLock redLock = new RedissonRedLock(lock1, lock2, lock3);
|
||||
|
||||
try {
|
||||
redLock.lock();
|
||||
// 业务逻辑
|
||||
} finally {
|
||||
redLock.unlock();
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
#### **坑 3：释放锁时误删**
|
||||
|
||||
**场景**：
|
||||
```
|
||||
线程 A 获取锁
|
||||
线程 A 执行时间过长，锁超时过期
|
||||
线程 B 获取锁
|
||||
线程 A 执行完毕，释放锁（删除了线程 B 的锁！）
|
||||
```
|
||||
|
||||
**解决方案：Lua 脚本 + 唯一标识**
|
||||
|
||||
```lua
|
||||
-- Lua 脚本（原子操作）
|
||||
if redis.call('get', KEYS[1]) == ARGV[1] then
|
||||
return redis.call('del', KEYS[1])
|
||||
else
|
||||
return 0
|
||||
end
|
||||
```
|
||||
|
||||
**Java 实现**：
|
||||
```java
|
||||
public boolean unlock(String key, String value) {
|
||||
String luaScript =
|
||||
"if redis.call('get', KEYS[1]) == ARGV[1] then " +
|
||||
" return redis.call('del', KEYS[1]) " +
|
||||
"else " +
|
||||
" return 0 " +
|
||||
"end";
|
||||
|
||||
DefaultRedisScript<Long> script = new DefaultRedisScript<>(luaScript, Long.class);
|
||||
Long result = redisTemplate.execute(script, Collections.singletonList(key), value);
|
||||
|
||||
return result != null && result == 1;
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 4. Zookeeper 分布式锁
|
||||
|
||||
#### **原理**
|
||||
|
||||
利用 Zookeeper 的**临时顺序节点**：
|
||||
|
||||
```
|
||||
/lock
|
||||
/lock/node-0000000001 (客户端 1)
|
||||
/lock/node-0000000002 (客户端 2)
|
||||
/lock/node-0000000003 (客户端 3)
|
||||
```
|
||||
|
||||
**流程**：
|
||||
1. 客户端创建临时顺序节点
|
||||
2. 获取所有子节点，判断自己是否是序号最小的
|
||||
3. 如果是最小节点，获取锁
|
||||
4. 如果不是，监听前一个节点的删除事件
|
||||
|
||||
---
|
||||
|
||||
#### **Curator 实现**
|
||||
|
||||
```java
|
||||
@Component
|
||||
public class ZkDistributedLock {
|
||||
|
||||
private final InterProcessMutex lock;
|
||||
|
||||
public ZkDistributedLock(CuratorFramework curatorFramework) {
|
||||
this.lock = new InterProcessMutex(curatorFramework, "/lock");
|
||||
}
|
||||
|
||||
public void acquire() {
|
||||
try {
|
||||
lock.acquire();
|
||||
} catch (Exception e) {
|
||||
throw new RuntimeException("获取锁失败", e);
|
||||
}
|
||||
}
|
||||
|
||||
public void release() {
|
||||
try {
|
||||
lock.release();
|
||||
} catch (Exception e) {
|
||||
throw new RuntimeException("释放锁失败", e);
|
||||
}
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
**使用**：
|
||||
```java
|
||||
@Service
|
||||
public class OrderService {
|
||||
|
||||
@Autowired
|
||||
private ZkDistributedLock lock;
|
||||
|
||||
public void createOrder(Order order) {
|
||||
try {
|
||||
lock.acquire();
|
||||
|
||||
// 业务逻辑
|
||||
|
||||
} finally {
|
||||
lock.release();
|
||||
}
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 5. Redis vs Zookeeper
|
||||
|
||||
| 特性 | Redis | Zookeeper |
|
||||
|------|-------|-----------|
|
||||
| **性能** | 高（内存操作） | 低（磁盘 + ZAB 协议） |
|
||||
| **可靠性** | 中（可能丢锁） | 高（CP 一致性） |
|
||||
| **实现复杂度** | 简单 | 复杂 |
|
||||
| **获取锁方式** | 轮询 | Watcher 通知（事件驱动） |
|
||||
| **锁释放** | 超时自动释放 | 会话结束自动释放 |
|
||||
| **适用场景** | 高并发、对一致性要求不高 | 严格一致性要求 |
|
||||
|
||||
---
|
||||
|
||||
### 6. Redisson 原理
|
||||
|
||||
#### **看门狗（自动续期）**
|
||||
|
||||
```java
|
||||
RLock lock = redisson.getLock("myLock");
|
||||
lock.lock(); // 默认 leaseTime = -1（启用看门狗）
|
||||
```
|
||||
|
||||
**原理**：
|
||||
```
|
||||
1. 加锁成功，启动看门狗线程
|
||||
2. 看门狗每 10 秒检查一次
|
||||
3. 如果锁还存在，续期 30 秒
|
||||
4. 客户端宕机，会话结束，看门狗停止，锁自动过期
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
#### **公平锁**
|
||||
|
||||
```java
|
||||
RLock fairLock = redisson.getFairLock("myLock");
|
||||
fairLock.lock();
|
||||
```
|
||||
|
||||
**原理**：
|
||||
- 加锁时，添加到队列尾部
|
||||
- 按照请求顺序获取锁
|
||||
- 性能比非公平锁低
|
||||
|
||||
---
|
||||
|
||||
#### **读写锁**
|
||||
|
||||
```java
|
||||
RReadWriteLock rwLock = redisson.getReadWriteLock("myLock");
|
||||
rwLock.readLock().lock(); // 读锁（共享）
|
||||
rwLock.writeLock().lock(); // 写锁（独占）
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 7. 实际项目应用
|
||||
|
||||
#### **场景 1：秒杀系统**
|
||||
|
||||
```java
|
||||
@Service
|
||||
public class SeckillService {
|
||||
|
||||
@Autowired
|
||||
private RedissonClient redisson;
|
||||
|
||||
public void seckill(Long userId, Long productId) {
|
||||
RLock lock = redisson.getLock("seckill:" + productId);
|
||||
|
||||
try {
|
||||
// 尝试加锁（最多等待 3 秒，锁自动释放时间 10 秒）
|
||||
boolean locked = lock.tryLock(3, 10, TimeUnit.SECONDS);
|
||||
|
||||
if (!locked) {
|
||||
throw new BusinessException("抢购人数过多，请稍后再试");
|
||||
}
|
||||
|
||||
// 1. 查询库存
|
||||
Integer stock = redisTemplate.opsForValue().get("stock:" + productId);
|
||||
if (stock == null || stock <= 0) {
|
||||
throw new BusinessException("库存不足");
|
||||
}
|
||||
|
||||
// 2. 扣减库存
|
||||
redisTemplate.opsForValue().decrement("stock:" + productId);
|
||||
|
||||
// 3. 创建订单
|
||||
createOrder(userId, productId);
|
||||
|
||||
} catch (InterruptedException e) {
|
||||
Thread.currentThread().interrupt();
|
||||
throw new BusinessException("系统异常");
|
||||
} finally {
|
||||
if (lock.isHeldByCurrentThread()) {
|
||||
lock.unlock();
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
#### **场景 2：定时任务分布式锁**
|
||||
|
||||
```java
|
||||
@Component
|
||||
public class ScheduledTask {
|
||||
|
||||
@Autowired
|
||||
private RedissonClient redisson;
|
||||
|
||||
@Scheduled(cron = "0 */5 * * * ?") // 每 5 分钟
|
||||
public void execute() {
|
||||
RLock lock = redisson.getLock("scheduled-task");
|
||||
|
||||
try {
|
||||
// 只有一个实例执行
|
||||
boolean locked = lock.tryLock(0, 5, TimeUnit.MINUTES);
|
||||
|
||||
if (!locked) {
|
||||
return; // 其他实例正在执行
|
||||
}
|
||||
|
||||
// 执行任务
|
||||
doTask();
|
||||
|
||||
} catch (InterruptedException e) {
|
||||
Thread.currentThread().interrupt();
|
||||
} finally {
|
||||
if (lock.isHeldByCurrentThread()) {
|
||||
lock.unlock();
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
### 8. 阿里 P7 加分项
|
||||
|
||||
**深度理解**：
|
||||
- 理解 CAP 理论在分布式锁中的体现
|
||||
- 理解 Redis 的 ZSet 实现延迟队列的原理
|
||||
- 理解 ZAB 协议和 ZK 的一致性保证
|
||||
|
||||
**实战经验**：
|
||||
- 有处理分布式锁死锁的经验
|
||||
- 有分布式锁性能优化的经验
|
||||
- 有分布式锁监控的经验
|
||||
|
||||
**架构能力**：
|
||||
- 能设计高可用的分布式锁方案
|
||||
- 能设计分布式锁的容灾方案
|
||||
- 能设计分布式锁的降级方案
|
||||
|
||||
**技术选型**：
|
||||
- 了解 Redis、Zookeeper、Etcd 等多种分布式锁实现
|
||||
- 能根据业务特点选择合适的方案
|
||||
- 有自研分布式锁的经验
|
||||
Reference in New Issue
Block a user