Redis expiration policy

1) Timed expiration

A timer is set for each created KEY, and when it expires, the KEY is deleted.

Advantages: Periodically delete data, memory friendlyDisadvantage: Takes a lot of CPU to clear data, affects cache response time and throughput

2) Lazy Expiration

When accessing the KEY, it is determined whether it has expired, and the data will be cleared when it expires.

Advantage: save CPU resourcesCons: Not memory friendly

3) Periodically expires

At regular intervals, the expired keys in the memory are randomly cleared.
It is a compromise between timed expiration and lazy expiration.
Both lazy expiration and periodic expiration are used in Redis.

redis' memory elimination strategy

Pre-redis4.0 elimination strategy

volatile-lru: When the memory is not enough to accommodate newly written data, use the LRU (least recently used) algorithm to eliminate from the key with the expiration time set;allkeys-lru: Use the LRU (least recently used) algorithm to evict all keys when the memory is not enough for newly written data.volatile-random: When the memory is not enough to accommodate the newly written data, the data is randomly eliminated from the key with the expiration time set;allkeys-random: Randomly evict data from all keys when memory is insufficient for newly written data.volatile-ttl: When the memory is not enough to accommodate the newly written data, in the key with the expiration time set, it will be eliminated according to the expiration time, and the earlier expired will be eliminated first;noeviction: The default policy, when the memory is not enough to accommodate the newly written data, the new write operation will report an error.

New elimination strategy after redis4.0

volatile-lfu: Added in version 4.0, when the memory is not enough to accommodate newly written data, the LFU algorithm is used to delete keys in expired keys.allkeys-lfu: Added in version 4.0, when the memory is not enough to accommodate newly written data, the LFU algorithm is used to eliminate all keys;