The first day of reading mysql45

The permission verification of the connector part is just whether the person has permission to log in ？ It involves whether the access user has permission to access sql Table used in ？
Query cache part , How to judge what to do sql It's in the cache ？sql Save all the statements and compare them ？
What does the optimizer optimize ？ and sql What is the difference between optimization ？ Be similar to java Instruction rearrangement in ？
The actuator operates the engine , About how to operate ？
In what form is the data stored in the storage engine ？

Broadly ,mysql Can be divided into server Layer and engine layer

mysql There are many cryptographic functions in , however mysql8 Many encryption functions have been removed from .

Different storage engines' tables access data in different ways , Functions are also different , When creating tables , use engine=InnoDB Specify the engine of the table , The default is InnoDB engine .
Before using the database, you need to log in to the database , This is the time to use The connector . The connector is responsible for establishing a connection with the client , Maintenance connection , Access permissions , Manage connections .

After success , The obtained permission will always be used in this connection , If the current connection is still valid , The user's permission is modified , You need to disconnect and reconnect to use the new permissions .

If after connection , No tasks performed , You'll be idle , Use show processlist Command to view the current and using database connections . The following screenshots ：

Command The value in the column is sleep It means it is idle , Idle time exceeds wait_timeout The connection will be automatically disconnected after time , You can see wait_timeout Configured time ：

But here 7200 I don't know what it is , Normally, it should be seconds , If it's minutes, it's 120 Hours ( It's not bullshit ), The unit should be seconds , That's it 2 Hours .

A long connection ： After the client connects to the database , The request will continue ; Short connection refers to the connection with the database , Idle after just a few requests .

But long connections can be expensive mysql The memory occupied by , because mysql The memory during execution is placed in the connection object , Only after the connection is disconnected , To clear the memory .

If the long connection persists , As a result, more and more memory will be used , Killed by the system . From a phenomenological point of view mysql Because of the abnormal restart .

How to solve the problem of mysql Restart problem ？
1. Regularly disconnect long connections , Free memory , Then reconnect .
2. It is possible to complete the execution of a project that consumes a lot of resources sql after , By calling mysql_reset_connection To reinitialize the connection resources , After calling , Will free up memory , The cache will also fail , Return to the state where the connection was started .

The frequency of query cache invalidation is very high , An update to a table will invalidate all cached records on the table , If a table is rarely updated , Insert, etc , Suitable for caching , Such as configuration tables .
Caching is based on query statements hash As key, Query results as value, With key-value Form to save , But if sql The case of the statement has changed , Or add notes , Lead to sql Of the statement hash The value has changed , Will result in no cache hit .
mysql8 Can be configured before query_cache_type Parameter disable query cache ,mysql8 The function of whole block query cache is deleted .
analyzer

A miss in the query cache will only , The parser will start to work . First of all “ Lexical analysis ”, It will write you select Identify this query keyword , Will write in from hinder week The identification is the table name
hold id Identify Columns id...; After lexical analysis of each keyword , It will be parsed , Judge whether there is a problem in grammar , Usually they meet “You have an error in your SQL syntax” Report errors . Namely
There is a problem in grammar .

Optimizer

After the analyzer ,mysql Will know you sql What exactly do you want to check , The optimizer optimization processing of the optimizer is also required after the actual execution . When the table has multiple indexes , The optimizer decides which indexes to use ,
Or there's a join The operation of the table , The optimizer will determine the priority according to what conditions to associate .

select  *
from week w
join week w2 on w2.id = w.id
where w.id = 2
and w2.id = 6

You can start with w.id = 2 Conditions to w Query the values in the table , Go back and w2 Tables are associated , Re execution w2.id=6, To w2 Find... In the table id be equal to 6 The data of ;
You can also start with w2.id=6 Go first to w2 Find... In the table id be equal to 6 Value , Go back and w Tables are associated , find w.id=2 Value .

Although the execution result is the same , But the execution efficiency may be different （ This is the same table used to save time , Different watches should be used ）. The role of the optimizer is to decide which execution scheme to adopt .

actuator

After going through the optimizer , The actuator is ready to start execution

select  *
from week w
where w.id  =6

Before starting the query , First, query whether you have permission for this table , Do not have permission to directly throw an error , If so, continue to carry out . Call the engine interface used to query .

1. call InnoDB The engine interface queries the first row of table data , Judge id Is the value equal to 6, If yes, put it in the result set , No, just skip

2. Call the engine interface to fetch the next row of data , Judge again

3. Execute until the last row of the table data , Return the data in the result set .

If the column to be queried is indexed , Would call InnoDB Engine “ Take the first line that satisfies the condition ” Interface , Then go to the loop call “ That satisfies the next row ”;