当前位置：网站首页>Summer Challenge database Xueba notes (Part 2)~

Summer Challenge database Xueba notes (Part 2)~

2022-07-07 02:24:00 【51CTO】

This article is participating in the starlight project 3.0– Summer Challenge

View / stored procedure / trigger View

View is a virtual table , Unlike tables that contain data , Views only contain queries that retrieve data dynamically when in use , It is mainly used to query . Why use views

       
        reusing sql sentence 
       
 Simplify complex sql operation , After writing the query , It can be easily reused without knowing its basic query details .
       
 Use parts of the table instead of the entire table .
       
 Protection data . Users can be granted access to specific parts of the table instead of the entire table .
       
 Change data format and presentation . Views can return data that is different from the representation and format of the underlying table .
      
1.
2.
3.
4.
5.

Be careful ：

       
        After view creation , They can be used in much the same way as tables . Can execute on view select operation , Filtering and sorting data , Join a view to another view or table , Can even add and update data .
       
 It's important to know that a view is just a facility for viewing data stored elsewhere . The view itself does not contain data , So the data they return is retrieved from other tables . When adding and changing data in these tables , The view will return the changed data .
       
 Because the view does not contain data , So every time you use a view , Must handle any retrieval required for query execution . If you use multiple joins and filters to create complex views or nested views , It may be found that the performance is greatly reduced . therefore , Before deploying an application that uses a large number of views , It should be tested .
      
1.
2.
3.

Rules and restrictions for views

       
        Like a watch , Views must be uniquely named ;
       
 You can create as many views as you want ;
       
 To create a view , You must have sufficient access rights . These restrictions are usually granted by the database administrator .
       
 Views can be nested , You can use queries that retrieve data from other views to construct a view .
       
Order by  Can be used in views , But if you retrieve data from this view select It also contains order by, So the order by  Will be overwritten .
       
 View cannot be indexed , You cannot have an associated trigger or default 
       
 Views can be used with tables 
      
1.
2.
3.
4.
5.
6.
7.

View creation

       
        utilize create view  Statement to create a view 
       
 Use show create view viewname; To view the statement that created the view 
       
 use drop view viewname  To delete the view 
       
 Update the view first drop stay create, You can also use create or replace view.
      
1.
2.
3.
4.

Update of view

Whether the view can be updated , It depends on the situation .

Usually the view can be updated , You can do insert,update and delete. Updating a view is updating its base table ( The view itself has no data ). If you add or delete rows to a view , In fact, it is to add or delete rows to the base table .

however , If MySQL Can't determine the updated base table data correctly , Update is not allowed ( Include insert and delete ), This means that the view cannot be updated if the following operations exist in the view ：(1) grouping ( Use group by and having );(2) coupling ;(3) Subquery ;(4) and ;(5) Aggregation function ;(6)dictinct;(7) export ( Calculation ) Column .

stored procedure

A stored procedure is one or more stored procedures for later use MySQL Collection of statements . It can be regarded as a batch document , Although their role is not limited to batch processing . Why use the storage process ？

1. By encapsulating the processing in an easy-to-use unit , Simplify complex operations ;

2. Since it is not required to repeatedly establish a series of processing steps , Data integrity guaranteed . If all developers and applications use the same ( Experiments and tests ) stored procedure , The code used is the same . The extension of this is to prevent mistakes . The more steps you need to take , The more likely it is to make a mistake , Error prevention ensures data consistency .

3. Simplify management of change , If table name . Column name or business logic changes , Just change the code of the stored procedure . People who use it don't even need to know about these changes . This extension is security , Limiting access to base data through stored procedures reduces the chance of data corruption .

4. Improve performance . Because using stored procedures is better than using separate sql Faster sentences .

5. There are some that can only be used for a single request MySQL Elements and properties , Stored procedures can use them to write more powerful and flexible code

Sum up ：

Three main benefits ： Simple 、 Security 、 High performance .

Two flaws ：

1、 The writing of stored procedures is more complicated , Need more skills, more experience .

2、 You may not have security access to create stored procedures . Many database administrators restrict stored procedures Create permissions , Allow to use , Creation of . Execute stored procedures

Call keyword ：Call Accept the name of the stored procedure and any parameters that need to be passed to it . Stored procedures can display results , You can also not show the results .

CREATE PROCEDURE productpricing()

       
       BEGIN
       
    SELECT  AVG( prod_price)  as priceaverage FROM products;
       
END;
      
1.
2.
3.
4.
5.

Create a productpricing The storage process of . If parameters need to be passed in the stored procedure , Then list them in brackets . Brackets must have .BEGIN and END The keyword is used to restrict the stored procedure body . The stored procedure body itself is a simple select sentence . Note that it's just to create the stored procedure without calling .

Use of storage process ：

Call productpring();

Stored procedure using parameters

General stored procedures do not show results , Instead, return the result to the variable you specified .

Variable ： A specific location in memory , Used to temporarily store data .

       
       MySQL> CREATE PROCEDURE prod(
       
     out pl decimal(8,2),
       
     out ph decimal(8,2),
       
     out pa decimal(8,2)
       
     )
       
begin
       
     select Min(prod_price) into pl from products;
       
     select MAx(prod_price) into ph from products;
       
     select avg(prod_price) into pa from products;
       
 end;
       
call PROCEDURE(@pricelow,@pricehigh,@priceaverage);
       
select @pricelow;
       
select @pricehigh;
       
select @pricelow,@pricehigh,@priceaverage;
      
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.

explain ：

This stored procedure accepts 3 Parameters ,pl Lowest price for storage products ,ph The highest price of storage products ,pa Average price of storage products . Each parameter must specify the type , The decimal system is used , keyword OUT Indicates that the corresponding parameter is used to transfer a value from the stored procedure ( Back to the caller ).

MySQL Support in( Pass to stored procedure )、out( Out of stored procedure , Used here ) and inout( Pass in and out of stored procedures ) Parameters of type . The code for the stored procedure is located in begin and end Statement within . They are a series of select sentence , Used to retrieve values . Then save to the corresponding variable ( adopt INTO keyword ).

Parameters of stored procedures allow the same data types as those used in tables . Note that recordsets are not allowed types , therefore , Cannot return more than one row and column with one parameter , That's why it's used 3 Parameters and 3 strip select The reason for the statement .

call ： To call this stored procedure , Must specify 3 Variable names . As shown above .3 Parameters are stored in the stored procedure 3 The names of variables . Invocation time , Statement does not display any data , It returns variables that can be displayed later ( Or in other processes ).

Be careful ： be-all MySQL Variables are all based on @ start .

       
       CREATE PROCEDURE ordertotal(
       
     IN innumber int,
       
     OUT outtotal decimal(8,2)
       
     )
       
     BEGIN
       
     SELECT Sum(item_price * quantity) FROM orderitems WHERE order_num = innumber INTO outtotal;
       
     end    //
       
CALL ordertotal(20005,@total);
       
select @total;  //  obtain 20005 Total of orders 
       
CALL ordertotal(20009,@total);
       
select @total; // obtain 20009 Total of orders 
      
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.

Stored procedures with control statements

 
CREATE PROCEDURE ordertotal(
       
 IN onumber INT,
       
 IN taxable BOOLEAN,
       
 OUT ototal DECIMAL(8,2)
       
 )COMMENT 'Obtain order total, optionally adding tax'
       
BEGIN
       
  -- declear variable for total
       
  DECLARE total DECIMAL(8,2);
       
  -- declear tax percentage
       
  DECLARE taxrate INT DEFAULT 6;
       
  -- get the order total
       
    SELECT Sum(item_price * quantity) FROM orderitems WHERE order_num = onumber INTO total;
       
-- IS this taxable?
       
IF taxable THEN
       
   -- yes ,so add taxrate to the total
       
   SELECT total+(total/100*taxrate)INTO total;
       
   END IF;
       
 -- finally ,save to out variable
       
 SELECT total INTO ototal;
       
END;
      
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.

We used... In the stored procedure DECLARE sentence , They define two local variables ,DECLARE Requires variable name and data type to be specified . It also supports optional defaults (taxrate Default 6%), Because later we have to decide whether to increase taxes , therefore , We put SELECT The results of the query are stored in local variables total in , And then in IF and THEN With the help of , Check taxable Is it true , And then in the real case , We use the other one SELECT Statement to increase business tax to local variable total in , Then we use SELECT Statement will total( The result of increasing or not increasing taxes ) Save to the general ototal in .

COMMENT keyword above COMMENT Can give or not give , If given , Will be in SHOW PROCEDURE STATUS The results show that .

trigger

Automatically process certain statements when a table changes , This is the trigger .

The trigger is MySQL Respond to delete 、update 、insert 、 be located begin and end An automatic execution of a set of statements between statements MySQL sentence . Other statements do not support triggers . Create trigger

When creating triggers , Need to give 4 statement （ The rules ）：

Unique trigger name ;
Trigger associated table ;
Trigger should respond to the activity ;
When the trigger executes ( Before or after handling )

Create trigger Sentence creation trigger

CREATE TRIGGER newproduct AFTER INSERT ON products FOR EACH ROW SELECT 'Product added' INTO @info;

CREATE TRIGGER Used to create a file named newproduct New trigger for . Triggers can be executed before or after an operation , here AFTER INSERT This trigger is in INSERT Statement executed after successful execution . This trigger also specifies FOR EACH ROW , So the code will execute on every insert line . Text Product added Will show once for each inserted row .

Be careful ：

1、 Triggers are only supported by tables , View , Triggers are not supported for temporary tables .

2、 Triggers are defined each time for each event in each table , Only one trigger per event per table is allowed at a time , therefore , Each table supports up to six triggers (insert,update,delete Of before and after).

3、 A single trigger cannot be associated with multiple events or tables , therefore , You need a right insert and update Trigger for operation execution , You should define two triggers .

4、 Trigger failed ： If before Trigger failed , be MySQL The requested operation will not be performed , Besides , If before Trigger or statement itself failed ,MySQL Will not execute after trigger . Trigger category

INSERT trigger

Is in insert Trigger executed before or after statement execution .

1、 stay insert Trigger code , One can be introduced called new The virtual table of , Access the inserted row ;

2、 stay before insert Trigger ,new Values in can also be updated ( Allow changes to inserted values );

3、 about auto_increment Column ,new stay insert Include... Before execution 0, stay insert Include new auto generated values after execution

CREATE TRIGGER neworder AFTER INSERT ON orders FOR EACH ROW SELECT NEW.order_num;

Create a file called neworder The trigger of , according to AFTER INSERT ON orders perform . Insert a new order into orders Table time ,MySQL Generate a new order number and save it to order_num in . Trigger from NEW.order_num Take this value and return it . This trigger must follow AFTER INSERT perform , Because in BEFORE INSERT Before statement execution , new order_num Not yet generated . about orders This trigger always returns a new order number for each insert of .

DELETE trigger

Delete Trigger in delete Statement before or after execution .

1、 stay delete Trigger code inside , You can quote a name as OLD The virtual table of , To access the deleted row .

2、OLD The values in are all read-only , Can't update .

       
       CREATE TRIGGER deleteorder BEFORE DELETE ON orders FOR EACH ROW
       
BEGIN
       
INSERT INTO archive_orders(order_num,order_date,cust_id) values (OLD.order_num,OLD.order_date,OLD.cust_id);
       
END;
       
----------------------------------------------------------------
       
CREATE TABLE archive_orders(
       
  order_num int(11) NOT NULL AUTO_INCREMENT,
       
  order_date datetime NOT NULL,
       
  cust_id int(11) NOT NULL,
       
  PRIMARY KEY (order_num),
       
  KEY fk_orders1_customers1 (cust_id),
       
  CONSTRAINT fk_orders1_customers1 FOREIGN KEY (cust_id) REFERENCES customers
       
 (cust_id)
       
) ENGINE=InnoDB AUTO_INCREMENT=20011 DEFAULT CHARSET=utf8
      
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.

This trigger will be executed before any order is deleted , It uses a INSERT Statement will OLD The value in ( Order to be deleted ) Save to a file named archive_orders In the archive form of ( For practical use of this example , We need to use with orders The same column is created with the name archive_orders Table of )

Use BEFORE DELETE Advantages of triggers ( be relative to AFTER DELETE The trigger says ) by , If for some reason , Orders cannot be filed ,delete Itself will be abandoned .

We used... In this trigger BEGIN and END Statement tag trigger body . This is not necessary in this example , Just to illustrate the use of BEGIN END The advantage of blocks is that triggers can hold multiple SQL sentence ( stay BEGIN END One by one in the block ).

UPDATE trigger

stay update Statement before or after execution

1、 stay update Trigger code inside , You can quote a name as OLD The virtual table of , Used to visit before (UPDATE The statement before ) Value , Quote a name as NEW Virtual table access new updated values .

2、 stay BEFORE UPDATE Trigger ,NEW The values in may also be used to update ( Allow changes to be used for UPDATE Value in statement )

3、OLD The values in are all read-only , Can't update .

CREATE TRIGGER updatevendor BEFORE UPDATE ON vendors FOR EACH ROW SET NEW.vend_state = Upper(NEW.vemd_state);

Make sure that state abbreviations are always capitalized ( No matter UPFATE Is capitalization given in the statement ), Every time a line is updated ,NEW.vend_state The value in ( Will be used to update the value of the table row ) Use both Upper(NEW.vend_state) Replace . summary

1、 Usually before For data validation and purification ( To ensure that the data inserted into the table is really the data needed ) Can also be applied to update trigger .

2、 And others DBMS comparison ,MySQL 5 The triggers supported in are fairly rudimentary , In the future MySQL It is estimated that there will be some improved and enhanced trigger support in the version .

3、 Creating triggers may require special security access , But the trigger executes automatically , If insert,update, perhaps delete Statement can execute , Then the relevant trigger can also execute .

4、 Use triggers to ensure data consistency ( Case write , Format, etc. ). The advantage of doing this type of processing in a trigger is that it always does it , And transparently , Nothing to do with client applications .

5、 A very meaningful use of triggers is to create audit trails . Use triggers , Change ( if necessary , Even before and after ) It's very easy to record to another table .

6、MySQL Trigger does not support call sentence , Cannot call stored procedure from within trigger .

Database recovery implementation technique

       
        Basic principles of recovery operation ： redundancy  
       
 Two key issues involved in the recovery mechanism 
       
 How to build redundant data 
       
     Data dump （backup）
       
     Log in log file （logging）
       
  How to use these redundant data to implement database recovery 
      
1.
2.
3.
4.
5.
6.
7.

Data dump

       
        Data dump definition ：
      
1.

Dump means DBA The process of copying the entire database to other storage media and saving it , Backup data is called a backup copy or backup copy

       
        How to use 
       
     After the database is destroyed, the backup copy can be reloaded 
       
     Reinstalling a backup copy can only restore the database to the state it was in when it was dumped 
       
 Dump method 
       
     Static dump and dynamic dump 
       
     Mass dump and incremental dump 
      
1.
2.
3.
4.
5.
6.

Static dump ：

1） Definition ： Dump operation when no transaction is running in the system . The database is in one at the beginning of the dump Sexual state , Dump does not allow any access to the database 、 To amend . Static dump must be a consistent copy of the data .

2） advantage ： Implement a simple

3） shortcoming ： Reduced database availability

The dump must wait for the end of a running user transaction to proceed ; New transactions must wait for the end of the dump to execute

Dynamic dump ：

       
               1） Definition ： Allow access to or modification of the database during dump . Dumps and user transactions can be executed concurrently .
       
        2） advantage ： Don't wait for a running user transaction to end ; Does not affect the operation of new transactions .
       
        3） Realization ： It is necessary to register the modification activities of each transaction to the database during the dump , Establishing a backup copy of the log file and adding the log file can restore the database to the correct state at a certain time .
      
1.
2.
3.
4.
5.

Mass dump ：

       
               1） Definition ： Dump all databases at a time 
       
        2） characteristic ： From the perspective of recovery , It is more convenient to use backup copies from mass dump for recovery .
      
1.
2.
3.

Incremental dump ：

       
               1） Definition ： Only the data updated after the last dump is dumped each time 
       
        2） characteristic ： If the database is large , Transaction processing is very frequent , Then incremental dump is more practical and effective .
      
1.
2.
3.

Log files

1、 What is a log file

Log files (log) It is used to record the update operation of transaction to database 2、 Format of log file

1） In records ：

The contents to be registered in the log file include ：

       
            Start mark of each transaction (BEGIN TRANSACTION)
       
     The end tag of each transaction (COMMIT or ROLLBACK)
       
     All update operations for each transaction 
       
  All of the above are recorded as a log in the log file                    
      
1.
2.
3.
4.
5.

What each log records ：

       
        Transaction ID （ Indicate which business ）
       
 Operation type （ Insert 、 Delete or modify ）
       
 Action object （ Record internal identification ）
       
 Old value of data before update （ For insertion operations , This item is null ）
       
 The new value of the updated data （ For delete operations ,  This item is null ）
      
1.
2.
3.
4.
5.

2） In blocks

The log records include ：

Transaction ID （ Indicate which business ）

Updated data block 3、 The role of log files

       
        Perform transaction recovery 
       
 Carry out system fault recovery 
       
 Assist backup copy in media recovery 
      
1.
2.
3.

1） Log files must be used for transaction recovery and system recovery

2） Log files must be created in dynamic dump mode , Only by combining backup copy with log files can the database be recovered effectively

3） Log files can also be created in static dump mode （ Reload backup copy , Then use the log file to redo the completed transaction , Undo outstanding transactions ）

4、 Register log files ：

       
        The basic principle of 
       
     The order of registration is strictly in the order of parallel transaction execution 
       
     You have to write a log file first , Write database after 
      
1.
2.
3.

Why write a log file first ？

1） Writing a database and a log file are two different operations , There may be a fault between these two operations

2） If the database modification is written first , This change is not registered in the log file , Then we can't restore this modification later

3） If you write a log first , But it didn't change the database , It is unnecessary to perform the log file recovery only once more UNDO operation , It doesn't affect the correctness of the database

Recovery strategy

Recovery of transaction failure

       
        Transaction failure ： Transaction terminated before running to normal termination 
       
 Recovery method 
       
     The recovery subsystem should use log files to undo （UNDO） Changes made to the database by this transaction 
       
 The recovery of transaction failure is automatically completed by the system , Transparent to users , No user intervention is required 
       
 Recovery steps for transaction failure 
      
1.
2.
3.
4.
5.

Reverse scan file log , Find the update operation of the transaction .
Reverse the update operation of the transaction . Is about to log “ Value before update ” Write to database .
The insert , “ Value before update ” It's empty , It is equivalent to deleting Delete operation ,“ Updated value ” It's empty , It's equivalent to inserting If it's a modification , It is equivalent to replacing the modified value with the pre modified value
Continue reverse scanning log files , Find other update operations for this transaction , And do the same thing .
Deal with it like this , Until you read the start tag of this transaction , Transaction recovery is complete . Recovery of system failure
The reason of database inconsistency caused by system failure Incomplete transaction update to database written to database The update of the committed transaction to the database is still left in the buffer before it can be written to the database Recovery method Undo Unfinished transactions at the time of failure Redo Completed transactions The recovery of system failure is automatically completed when the system is restarted , No user intervention is required Recovery steps for system failure
Forward scanning log files

       
             redo (REDO)  queue :  Transactions committed before the failure 
       
          These things already have BEGIN TRANSACTION Record , Also have COMMIT Record 
       
      revoke  (Undo) queue :  Unfinished business at the time of failure 
       
           These things are just BEGIN TRANSACTION Record , There is no corresponding COMMIT Record 
      
1.
2.
3.
4.

To revoke (Undo) Queue transactions to undo (UNDO) Handle

       
             Reverse scan log files , For each UNDO The update operation of the transaction performs the reverse operation 
      
1.

To redo (Redo) Queue transactions redo (REDO) Handle

       
             Forward scanning log files , For each REDO The transaction re performs the enlistment operation 
      
1.

Recovery of media failure

Recovery steps

Reload the database

       
        Mount the latest backup copy , Restore the database to the consistency state of the last dump .
       
 For a database copy of a static dump , The database is in a consistent state after loading 
       
 For a database copy of a dynamic dump , You must also load a copy of the log file at the time of the dump , Use the method of recovering system failure （ namely REDO+UNDO）, To restore the database to a consistent state .
      
1.
2.
3.

Load a copy of the log file in question , Redo what has been done .

       
        First scan the log file , Find out the ID of the transaction that was committed when the failure occurred , Put it in the redo queue .
       
 Then forward scan the log file , Redo all transactions in the redo queue .
       
 Recovery from media failure requires DBA intervention 
       
DBA The job of 
       
     Reinstall the database copy of the latest dump and the copies of the related log files 
       
     Execute the recovery command provided by the system , Specific recovery operations are still performed by DBMS complete 
      
1.
2.
3.
4.
5.
6.

checkpoint

       
        There are two problems in using log technology to recover database 
       
     Searching the entire log will take a lot of time 
       
    REDO Handle ： The transaction has actually been executed , And re execute , Wasted a lot of time 
       
 With checkpoints （checkpoint） Recovery technology 
       
     Add checkpoint records to the log file （checkpoint）
       
     Add restart file , And let the recovery subsystem dynamically maintain the log during log file login 
       
 What the checkpoint records 
       
     Create a list of all transactions that are executing at checkpoint time 
       
     The last logged address of these transactions 
       
 Restart the contents of the document 
       
     Record the address of each checkpoint recorded in the log file 
       
 How to maintain log files dynamically 
      
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.

Periodically do the following ： Set up checkpoints , Save database state .

The specific steps are ：

1. Write all the log records in the current log buffer to the log file on disk

2. Write a checkpoint record in the log file

3. Write all data records of the current data buffer to the database of the disk

4. Write the address of the checkpoint recorded in the log file to a restart file

Using checkpoint methods can improve recovery efficiency

When a transaction T Submit... Before a checkpoint ：

       
       T Changes to the database must have been written to the database 
       
 The write time is before or when the checkpoint is established 
       
 In recovery processing , There is no need to T perform REDO operation 
      
1.
2.
3.

Use checkpoint recovery steps

1. Find the address of the last checkpoint recorded in the log file... From the restart file , This address finds the last checkpoint record in the log file

2. The checkpoint records the list of all transactions being executed at the checkpoint establishment time ACTIVE-LIST

       
        Create two transaction queues 
       
    UNDO-LIST
       
    REDO-LIST
       
     hold ACTIVE-LIST Put in for a while UNDO-LIST queue ,REDO The queue is temporarily empty 
      
1.
2.
3.
4.

3. Scan log files forward from checkpoint , Until the end of the log file

       
        If there is a new beginning Ti, hold Ti Put in for a while UNDO-LIST queue 
       
 If there are committed transactions Tj, hold Tj from UNDO-LIST The queue moves to REDO-LIST queue 
      
1.
2.

4. Yes UNDO-LIST Each transaction in performs UNDO operation

       
                      Yes REDO-LIST Each transaction in performs REDO operation 
      
1.

Mirror image

In order to avoid hard disk media failure affecting the availability of the database , many DBMS Database image provided （mirror） Function for database recovery .

Copy the entire database or critical data from it to another disk , Whenever the primary database is updated ,DBMS Automatically copy the updated data , from DBMS Automatically ensure the consistency between the image data and the primary database . In case of media failure , It can be used by the mirror disk , meanwhile DBMS Automatic use of disk data for database recovery , There is no need to shut down the system and reinstall the database copy .

When there is no failure , Database mirroring can also be used for concurrent operations , That is, when a user adds an exclusive lock to the database to modify data , Other users can read the data on the mirror database , Instead of waiting for the user to release the lock .

Because database mirroring is achieved by copying data , Assigning data frequently will naturally reduce the efficiency of the system . Therefore, in practical applications, users often only choose to mirror key data and log files .

Summary ：

       
        If the database contains only the results of a successful transaction commit , Say the database is in a consistent state . To ensure data consistency is the most basic requirement of database .
       
 A transaction is the logical unit of work of a database 
       
    DBMS Ensure the atomicity of all transactions in the system 、 Uniformity 、 Isolation and continuity 
       
DBMS The transaction must fail 、 System failure and medium failure recovery 
       
 The most frequently used technology in recovery ： Database dump and register log files 
       
 The fundamentals of recovery ： Use the backup copy stored in 、 Log files and redundant data in the database image to rebuild the database 
      
1.
2.
3.
4.
5.
6.

Common recovery techniques

Recovery of transaction failure

Recovery of system failure

       
       UNDO + REDO
      
1.

Recovery of media failure

       
        Reinstall the backup and restore to a consistent state  + REDO
      
1.

Technology to improve recovery efficiency

Checkpoint Technology

       
        It can improve the recovery efficiency of system failure   
       
 To some extent, it can improve the efficiency of media recovery by dynamic dump backup 
      
1.
2.

mirror technique

       
        Image technology can improve the recovery efficiency of media failure 
      
1.

concurrency control

Multi user database ： Databases that allow multiple users to use at the same time （ Booking system ）

Different multi transaction execution methods ：

       
            1. Serial execution ： There is only one transaction running at a time , Other transactions must wait until the end of this transaction before running .
       
Cross concurrency ：
      




In a single processor system , The concurrent execution of transactions is actually the cross running of the parallel operations of these parallel transactions （ It's not really concurrency , But it improves the efficiency of the system ）

3. Concurrent mode ：

In a multiprocessor system , Each processor can run a transaction , Multiple processors can run multiple transactions at the same time , Realize the real parallel operation of multiple transactions

Problems caused by concurrent execution ：

       
             Multiple transactions access the same data at the same time （ Shared resources ）
       
      Access to incorrect data , Breaking transaction consistency and database consistency 
      
1.
2.
3.

summary

Data inconsistencies caused by concurrent operations include

       
            1） Missing changes （lost update）
       
     2） It can't be read repeatedly （non-repeatable read）
       
     3） Read dirty data （dirty read）
      
1.
2.
3.
4.
5.

mark :W（x） Writing data x R（x） Reading data x

The task of concurrency control mechanism ：

       
            1） Schedule concurrent operations correctly 
       
     2） Ensure the isolation of transactions 
       
     3） Ensure database consistency 
      
1.
2.
3.
4.
5.

The main technology of concurrency control

       
            1） The blockade （locking）（ Mainly used ）
       
     2） Time stamp （timestamp）
       
     3） Optimistic control （optimistic scheduler）
       
     4） Multi version concurrency control （multi-version concurrency control ,MVCC）
      
1.
2.
3.
4.
5.
6.
7.

The blockade

The blockade ： Blockade is business T On a data object （ Such as table 、 Records, etc. ） Before the operation , Make a request to the system first , Lock it . After lock transaction T It has certain control over the data object , In the transaction T Before releasing its lock , Other transactions cannot update this data object

The exact control depends on the type of blockade

There are two basic types of blockade ： Exclusive lock （X lock ,exclusive locks）、 Shared lock （S lock ,share locks）

Exclusive lock is also called write lock , Yes A With the lock on it , Nothing else can be right A Add Any type of lock （ Repel reading and writing ）

Shared lock is also called read lock , Yes A After adding the shared lock , Other things can only be done to A Add S lock , Cannot add X lock （ Write only ）

（ Very important ）

# Summer Challenge # Database Xueba notes （ Next ）~_ trigger

Blocking agreement

Using X Lock and S When locking data objects , Some rules need to be agreed ： Blocking agreement （Locking Protocol）

When to apply for X Lock or S lock 、 Holding time 、 When to release

Make different rules about how to block , There are different kinds of blockade agreements .

Common blocking protocol ： Three level blockade agreement

The three-level blocking protocol solves the concurrency problem in different degrees , It provides a guarantee for the correct scheduling of concurrent operations .

1、 First level blockade protocol

Business T Modifying data R Before , It must be added first X lock , Until the end of the transaction （commit/rollback） To release .

The first level blocking protocol can prevent the loss of modification

If it's reading data , It doesn't need to be locked , So it's not guaranteed to be repeatable or unreadable “ dirty ” data .

2、 Secondary blockade protocol

       
             On the basis of the first level blockade agreement （ Write with X lock , End of transaction release ） On , Add business T Reading in data R It must be added before S lock , After reading, you can release S lock .（ Read with S lock , Release after reading ）
      
1.

The second level blocking protocol can prevent the loss of modification , It also prevents reading dirty data

Because the data is released after reading S lock , There's no guarantee that it's not repeatable

3、 Three level blockade agreement ：

       
             Add transactions based on the level 1 blocking agreement T Reading data R It must be added before S lock , Until the end of the transaction .
      
1.

The three-level blocking protocol can prevent the loss of modification and reading of dirty data , It also prevents non repeatable reading

The main difference of the three-level blocking protocol is what operation needs to apply for a lock , When to release the lock . The higher the blockade , The more consistent .

hunger

hunger ： Business T1 Blocked data R, Business T2 And ask for a blockade R, therefore T2 wait for .T3 And ask for a blockade R, When T1 The release of the R After the blockade on , The system first approved T3 Request ,T2 Still waiting . T4 And ask for a blockade R, When T3 The release of the R After the blockade, the system approved T4 Request ……T2 It's possible to wait forever , This is the case of hunger

The way to avoid hunger ： First come, first served

When multiple transaction requests block the same data object , These transactions are queued in the order in which they are requested to be blocked

Once the lock on the data object is released , First approve the first transaction in the application queue to obtain the lock .

Deadlock

Deadlock ： Business T1 Blocked data R1, T2 Blocked data R2. T1 And ask for a blockade R2, because T2 It's blocked R2, therefore T1 wait for T2 Release R2 The lock on the . next T2 And apply for a blockade R1, because T1 It's blocked R1,T2 Only

wait for T1 Release R1 The lock on the . such T1 Waiting for the T2, and T2 Waiting again T1,T1 and T2 Two things can never end , Formation of a deadlock .

The solution to deadlock ： The prevention of 、 Diagnosis and removal

1、 Deadlock prevention

The cause of deadlock is that two or more transactions have blocked some data objects , Then they all request to lock data objects that have been blocked by other transactions , So there's a dead wait .

To prevent deadlock is to destroy the condition of deadlock

Method

1） One time blockade ：

       
             Each transaction is required to lock all data to be used at one time , Otherwise, we can't carry on .
       
      The problem is ： Reduce the concurrency of the system ; It is difficult to accurately determine the blocked object 
      
1.
2.
3.

2） Sequential blockade ：

       
             Set a blocking order for data objects in advance , All transactions are blocked in this order .
       
      The problem is ：
      
1.
2.
3.

Maintenance cost ： There are many blocked objects in the database system , And it's changing

Difficult to achieve ： It's hard to determine which objects to block for each transaction

DBMS A common method of diagnosing and unlocking

2、 Diagnosis and removal of deadlock

       
             Method ： Timeout method and transaction wait graph method 
       
     1） Overtime method ： If the waiting time of a transaction exceeds the specified time limit , Think of a deadlock 
       
               advantage ： Implement a simple 
       
               shortcoming ： Misjudgment of deadlock ; If the time limit is set too long , It can't be found in time after deadlock .
       
     2） Transaction wait graph ： Use the transaction waiting graph to dynamically reflect the waiting status of all transactions 
      
1.
2.
3.
4.
5.
6.
7.
8.
9.

Waiting graph is a directed graph G=(T,U),T For the set of nodes , Each node represents a running transaction , U A collection of sides , Each side represents the transaction waiting . if T1 wait for T2, be T1、T2 Draw a directed edge between , from T1 Point to T2.

The concurrency control subsystem periodically （ For example, every few seconds ） Generate transaction wait graph , Detect transactions . If a circuit is found in the diagram , It means there is a deadlock in the system .

Deadlock Relieving ： The concurrency control subsystem chooses a transaction with the least deadlock cost , Withdraw it .

Release all locks held by the transaction , Enable other transactions to continue .

Serial scheduling

What kind of scheduling is right ？ Serial scheduling is correct .

（ It is also true that the execution result is equivalent to serial scheduling , Such scheduling is called serializable scheduling .） Serializable scheduling

Definition ： Concurrent execution of multiple transactions is correct , If and only if the result is the same as when these transactions are executed serially in a certain order , This scheduling strategy is called serializable scheduling （serializable）.

Serializability is the criterion for the correct scheduling of concurrent transactions . According to this rule , A given concurrent schedule , If and only if it is serializable , I think it's the right scheduling .

Conflicts can be scheduled serially

Determine the sufficient conditions for serializable scheduling

Conflicting operations ： Different transactions read and write the same data .

Conflicting operations of different transactions and two operations of the same transaction cannot be exchanged .

Ri（x） and Wj（x） Not exchangeable ,Wi（x） and Wj（x） Not exchangeable

Conflicts can be scheduled serially ：

A dispatch Sc In the case of keeping the order of conflicting operations unchanged , Get another schedule by exchanging the order of two non conflicting transactions Sc’, If Sc’ It's serial , Call scheduling Sc Serializable scheduling for conflicts .

Two stage lock protocol

DBMS The concurrency control mechanism must provide some means to ensure that the scheduling is serializable . at present DBMS Two stage lock protocol is widely used （TwoPhase Locking, abbreviation 2PL） To show the serializability of concurrent scheduling .

Two phase locking protocol means that all transactions must be locked and unlocked in two phases .

       
            1） Before reading or writing any data , First of all, you need to apply for and obtain a lock on the data .
       
     2） After releasing a lock , The transaction no longer requests and obtains any other locks .
      
1.
2.
3.

“ Two paragraphs ” The meaning of lock ： The transaction is divided into two stages

The first stage is to get the blockade , Also known as the extension phase

A transaction can request any type of lock on any data object , But you can't release any locks

The second stage is to release the blockade , Also known as the contraction phase

A transaction can release any type of lock on any data object , But you can't apply for any more locks

It is a sufficient condition for serializable scheduling that transactions obey two-stage locking protocol , It's not necessary .

If both concurrent transactions follow the two-stage locking protocol , Any concurrent scheduling strategy for these transactions is serializable

If a schedule of concurrent transactions is serializable , Not all transactions are subject to the two-stage locking protocol

Two stage locking protocol and one-time blocking method to prevent deadlock

One time blocking requires every transaction to lock all data to be used at one time , Otherwise, we can't carry on , Therefore, the one-time blocking law abides by the two-stage locking agreement

However, the two-stage locking protocol does not require the transaction to lock all data to be used at one time , Therefore, transactions complying with the two-stage locking protocol may deadlock

The granularity of the blockade

The size of the blocking object is called blocking granularity （granularity）.

The blocked object can be a logical unit （ Property value 、 Property value set 、 Tuples 、 Relationship 、 Index entry 、 database ）, It can also be a physical unit （ page 、 Physical records ）.

Choose the principle of blocking granularity ：

       
             The granularity of blocking is closely related to the concurrency of the system and the overhead of concurrency control 
       
      The greater the granularity of the blockade , The fewer data units a database can block , The lower the concurrency , System overhead 
      
1.
2.
3.

The smaller it is ;

The smaller the size of the blockade , High concurrency , But the more overhead there is

Intent locks

Intent locks ： If an intention lock is applied to a node , The lower level node of the node is being locked ; When locking any node , You must first add intention lock to its upper node .

for example , When locking any tuple , You must first add an intentional lock to the database and relationship it is in .

Three kinds of common intention locks ： Intention sharing lock （Intent Share Lock,IS lock ）; Intention exclusive lock （Intent Exclusive Lock,IX lock ）; Sharing intention exclusive lock （Share Intent Exclusive Lock,SIX lock ）.

1、IS lock

If you add... To a data object IS lock , Indicates that its child nodes are to be added S lock .

for example ： Business T1 Right R1 Add... To a tuple in S lock , First of all, the relationship R1 And database plus IS lock

2、IX lock

If you add... To a data object IX lock , Indicates that its child nodes are to be added X lock .

for example ： Business T1 Right R1 Add... To a tuple in X lock , First of all, the relationship R1 And database plus IX lock

3、SIX lock

If you add... To a data object SIX lock , To add to it S lock , add IX lock , namely SIX = S + IX.

for example ： Add... To a watch SIX lock , It means that the transaction needs to read the whole table （ So add S lock ）, Same as

Will update individual tuples （ So add IX lock ）

Intention lock strength ： The strength of a lock is its rejection of other locks . It is safe for a transaction to replace weak lock with strong lock when applying for blocking , Otherwise .

Multi granularity blocking method with intention lock

When applying for a blockade, it should be done in a top-down order

The release of the blockade should be done in a bottom-up order

advantage ：

       
            1） Improve the system concurrency 
       
     2） Reduce the cost of locking and unlocking 
      
1.
2.
3.

In practice DBMS Widely used in products .

Other concurrency controls

Besides blocking technology, the method of concurrency control , And the time stamp method 、 Optimistic control and multi version concurrent control .

Timestamp method ： Put a time mark on each transaction , That is, the time when the transaction starts . Each transaction has a unique timestamp , And according to this timestamp to solve the conflict operation of the transaction . In case of conflicting operations , Roll back to a transaction with an earlier timestamp , To ensure the normal execution of other affairs , The transaction being rolled back is given a new timestamp and executed from the beginning .

Optimistic control law holds that there are few conflicts in transaction execution , So there's no special control over things , It's about letting it go , Check the correctness of the transaction before committing . If after checking, it is found that there are conflicts in the execution of the transaction and the serializability is affected , Then reject the commit and roll back the transaction . Also known as verification method

Multi version control is a strategy to achieve high efficiency and concurrency by maintaining multiple version information of data objects in database .