redis双写一致性旁路缓存

### Redis 双写一致性旁路缓存实现方案 #### 1. 缓存读取逻辑 当应用程序尝试获取某个键的数据时，会首先访问缓存。如果缓存命中，则直接返回缓存中的数据[^4]。 ```python def get_data_from_cache(key): data = redis.get(key) if data is not None: return data # Cache miss, read from database and populate cache data = db.query(key) # Check deleteKey to avoid dirty reads during replication delay if check_delete_key(key): data = master_db.query(key) # Force read from master else: data = slave_db.query(key) # Normal read from slave redis.setex(key, ttl, data) # Populate the cache with fresh data return data ``` #### 2. 数据更新逻辑 对于数据更新操作，在修改数据库之前先清除对应的缓存条目，并设置一个`deleteKey`标志位用于处理主从延迟期间可能发生的脏读问题[^3]。 ```python def update_data_in_database(key, value): set_delete_key(key) # Set a flag indicating this key was deleted recently try: result = db.update(key, value) clear_delete_key_after_delay(key) # Clear after estimated replication lag time return result finally: redis.delete(key) # Invalidate stale cached entry immediately ``` #### 3. 处理主从复制延迟 为了避免因MySQL主从同步延迟而导致的脏读现象，可以在删除缓存的同时创建一条带有过期时间(`TTL`)等于预计最大主从延迟周期的特殊标记记录(deleteKey)[^3]。 ```python import threading def set_delete_key(key): """Set a temporary marker that indicates recent deletion.""" redis.set(f'delete:{key}', 'true', ex=replication_lag_time) def check_delete_key(key): """Check whether there's an active delete mark for given key.""" return bool(redis.exists(f'delete:{key}')) def clear_delete_key_after_delay(key): timer = threading.Timer(replication_lag_time, lambda: redis.delete(f'delete:{key}')) timer.start() ``` 通过上述措施可以有效解决Redis作为旁路缓存时可能出现的一致性问题，既保持了良好的性能又确保了数据准确性。