mysql Master-slave replication of

1、 Main library Master node , Accept additions from the client 、 Delete 、 Change request , adopt IO-thread Write to binlog

2、 Slave Library slave node , Request to read the binlog, adopt IO-thread Write local relay log（ relay logs ）

3、 Pass through the library sql-thread Read relay-log, And implement the additions, deletions and modifications .

mysql Separation of reading and writing

among , Read by reading from the node , Writes are written through the master node and synchronized to the slave node , Read and write separation is achieved

Why mysql Caching scheme for ？

1、 Read more and write less , The access speed of memory is the same as that of disk 10 ten thousandfold （ Order of magnitude magnification ）, Memory access speed ⼤ About 100ns,⽽⼀ Secondary disk access ⼤ About 10ms; visit mysql The number of times the disk is accessed during is the same as b+ Treelike ⾼ Degree correlation ;
2. ⼀ like ⼤ Some items ⽬ in , Database read operations are write operations 10 About times ;
The buffer layer plays a good role in improving the reading performance

Cache layer scheme

Use hot data redis Cache and backup the database , Transfer hot read operations to the cache database

It can solve the problem of reading performance .

The cache layer is introduced , How to solve the problem of consistency ？

From the following situations ：

1、mysql No, , The cache has

2、mysql Yes , The cache has , But inconsistent

3、mysql Yes , The cache does not

According to business scenarios , Look at our requirements for consistency ：

The final consistency solution

If the project wants to achieve final consistency , Our solutions ：

read ： Read cache first , Cache has return , If not, read mysql, To write redis

Write ：1、 direct writing mysql, wait for mysql Synchronize to redis

       2、 Write cache first redis, Set timeout , To write mysql, Final mysql Synchronize to redis.

Strong consistency solutions

read ： Read cache first , There is a direct return to the cache , Cache does not exist , visit mysql obtain , To write redis

Write ： So let's delete the cache , To write mysql, wait for mysql Synchronize to cache .

The plan above , Assume ：mysql The final data source .

Synchronization scheme

git clone https://gitee.com/mirrors/go-mysql-transfer.git

go-mysql-transfer Realize to mysql Data synchronization to redis