How redis ensures data consistency

Nowadays, redis is widely used in the Internet. Its use varies depending on the business. However, there are generally problems with the consistency of data storage and caching. Here we will introduce to you some mature and practical solutions. .

Method 1:

Database saves data, redis does not persist After redis is started, loading data from the database is not required Read requests that require strong consistency and real-time performance are processed by redis. Read requests that require strong consistency and real-time performance are processed by the database. There are two processing methods for write requests, which are processed by the database. (Recommended learning: Redis Video Tutorial)

The application writes to the database first, and then updates redis

The application writes to the database first, and then other daemons synchronize to redis

Advantages: redis startup does not need to deal with inconsistencies between redis data and the database

Disadvantages: redis startup puts a lot of reading pressure on the database

Method 2:

Database and redis handle different data types respectively, database processing requirements Strongly consistent and real-time data, such as financial data and transaction data; Redis processes data that does not require strong consistency and real-time performance, such as the most popular post rankings on the website

The synchronization of redis and MySQL data, the code level can be roughly as follows Do: Read: Read redis->No, read mysql->Write mysql data back to redis Write: Write mysql->Success, write redis

Concurrency is not high: Read: Read redis- >No, read mysql->Write the mysql data back to redis, if there is, get it directly from redis; Write: write mysql->Success, then write to redis;

High concurrency situation: read : Read redis->No, read mysql->Write mysql data back to redis, if there is any, take it directly from redis; Write: If it is asynchronous, write it to the redis cache first and then return directly; regular or specific actions will Data is saved to mysql, which can be updated multiple times and saved once;

--Note: If you write redis, you need to use redis transactions:

127.0.0.1:6379> WATCH id
OK
127.0.0.1:6379> MULTI
OK
127.0.0.1:6379> INCR id
QUEUED
127.0.0.1:6379> EXEC
1) (integer) 342183
127.0.0.1:6379>

Method 3:

Use lua script: redis only allows one script to be executed at the same time when using lua, which is consistent with the atomicity of transactions, but a lua script cannot be executed for too long, otherwise it will block

EVAL

The EVAL command executes and evaluates Lua scripts.

Grammar:

EVAL script numkeys key [key …] arg [arg …]

script lua脚本内容 注意的是脚本不应该是Lua函数。
numkeys 表示指定键名参数的个数。
key [key ...] 表示脚本对应的key值列表 在脚本中可以使用KEYS[1] KEYS[2] KEYS[3]KEYS[n] n从1开始 。
arg [arg ...] 命名行中传递的参数列表 在脚本中可以使用ARGV[1] ARGV[2] ARGV[3]ARGV[n] n从1开始 。

An example is worth a thousand words of explanation

eval "return {KEYS[1],KEYS[2],ARGV[1],ARGV[2]}" 2 id name 3 mytest

Execute the above script to return

1) "id"2) "name"
3) "2"
4) "mytest"

For more Redis-related technical articles, please visit the Redis Getting Started Tutorial column to learn!

The above is the detailed content of How redis ensures data consistency. For more information, please follow other related articles on the PHP Chinese website!