Database optimization

1、SQL And index optimization

（ Large amount of data , perform SQL Low efficiency , To optimize ）

Open slow query log -》 Set a reasonable slow query time for the corresponding amount of item table data –》 Find record slow query SQL–》 use explain analysis SQL Implementation plan of –》 The corresponding optimization measures are given –》 If the amount of data is too large , Paging with . No index . The first column is not used in multi column indexing . Size table settings for federated queries . Multi table query does not need in, Replace the belt with an external connection in The process of subquery , Use index reasonably , No intermediate table will be generated .

2、 Application optimization

• 1、 Connect to database , Connection pool can be imported （ middleware ）
• 2、 Introduce caching （ Store hot data ,redis, If the client's request comes , First in redis Check it out , If redis If you find it directly on the database, you don't go through the database , If you don't find it, visit MySQL database , Visit back , First, send the accessed data to redis Let's cache it , And then return the results to the user .redis It's a memory based database ）
  Increase in business redis、memcache, Cache is generally used to cache frequently accessed data .
  introduce redis Cache words , It is necessary to prevent these problems ： Cache data consistency problem . Cache penetration and cache avalanche .

Besides optimization SQL And index , A lot of times , In the actual production environment , Due to the performance limitations of the database server itself , We must optimize the application of the upper layer , So that the pressure of the upper application accessing the database can be minimized .

Connection pool

Generally, the application accesses the database , All first MySQL Server Create connection , And then send SQL sentence ,Server After processing , Then return the result to the application through the network , Then close and MySQL Server The connection of , Therefore, a large number of database accesses in a short time , The consumption of TCP Three handshakes and four waves take a lot of time , A slightly larger project , We will all be on the layer where the application accesses the database , Add connection pool module , Equivalent to application and MySQL Server Create a set of connections in advance , When the application needs to request MySQL Server when , There's no need to go on TCP Connect and release the connection , Generally, the connection pool will maintain the following resources ：
1、 A fixed number of connections remain active in the pool TCP Connect , Supply and use .
2、 If the application has instant access MySQL The amount of is relatively large , Then the connection pool will create more connections for applications in real time .
3、 When in the connection pool TCP The connection has not been used for some time , The connection pool will free up redundant connection resources , Just keep the maximum number of free connections it sets .
The connection pool can be implemented by itself , You can also use a library written by a third party .

MySQL The query cache

MySQL The query cache is to put select Record the last query result of the query statement and put it in the cache , The next time you query the same content , Just take it out of the cache , You don't have to do it again SQL Inquire about . （ Divide a space in memory for cache ）

But when two select Appear in the middle of query insert,update,delete At the time of statement , The query cache will be emptied . Query caching is suitable for tables that are not updated frequently , Frequently queried tables , Because when the table is updated frequently , The query cache is always emptied , Too many queries add and delete cached data , It will affect MySQL Efficiency of execution , It's better to check it from the disk every time （ Cache refers to a piece of memory , Memory I/O Compare disk I/O Much faster ）.

Indexing and data caching

It mainly refers to innodb_buffer_pool_size Configuration item , You can see from the name , This configuration item is for InnoDB The storage engine works , This parameter defines InnoDB The maximum memory buffer size for storing the engine's table data and index data .
innodb_buffer_pool_size Is to cache data blocks and index blocks at the same time , The higher this value is set , The disks needed to access the data in the table I/O The less .

MySQL Thread cache

It mainly refers to... In the configuration file thread_cache_size Configuration item . I told you MySQL Server The network module adopts the classic I/O Reuse + Thread pool model , The reason why thread pool is introduced , Mainly for the purpose of business execution , Not because of temporary creation and destruction of threads , Cause system performance degradation , Because the creation and destruction of threads is very performance consuming , So thread pool is before business use , First create a fixed number of threads , Waiting for events to happen , When there is SQL Request arrival MySQL Server When , Take a thread from the thread pool to execute the process SQL Just ask , After execution , Don't destroy threads , Instead, return the thread to the thread pool , Wait for the processing of the next task （MySQL According to the number of connections , Automatically increase the number of thread pools ）.（ The number of threads in the thread pool increases with requests , It can be added dynamically ）
Scalable thread pool （ Initialization is 1 Threads , Handle all business , If the business request cannot be processed , Thread requests to the thread pool always fail , At this point, the thread pool will create more threads to process more requests , Some time after processing , Maybe there are not so many requests , Idle threads will return to the thread pool , The thread pool will also reduce the number of threads in the thread pool according to the corresponding configuration , Return resources to the system ）

3、MySQL Server Optimize

Configuration of various parameters

Turn off adaptive hash indexing , Maintaining an adaptive hash index takes a lot of performance , If you rarely use , Turn it off through parameter configuration .

Improve redo log performance , Can be reset InnoDB log buffer size （redo log Cache size ）,Innodb_buffer_pool_size（ Cache size ）, To reduce disk I/O frequency , Because the cache is large , Working in the buffer for a long time , It's more efficient .

about MySQL Server End optimization , It mainly refers to MySQL Server Optimization of configuration item content of configuration file loaded at startup