Mysql加鎖過程詳解（3）-關於mysql 幻讀理解

Mysql加鎖過程詳解（1）-基本知識

Mysql加鎖過程詳解（2）-關於mysql 幻讀理解

Mysql加鎖過程詳解（3）-關於mysql 幻讀理解

Mysql加鎖過程詳解（4）-select for update/lock in share mode 對事務並發性影響

Mysql加鎖過程詳解（5）-innodb 多版本並發控制原理詳解

Mysql加鎖過程詳解（6）-數據庫隔離級別（1）

Mysql加鎖過程詳解（6）-數據庫隔離級別（2）-通過例子理解事務的4種隔離級別

Mysql加鎖過程詳解（7）-初步理解MySQL的gap鎖

Mysql加鎖過程詳解（8）-理解innodb的鎖(record,gap,Next-Key lock)

 

 

Mysql加鎖過程詳解（2）-關於mysql 幻讀理解出現了幻讀，那么不是說mysql的重復讀解決了幻讀的么？

 

那么，InnoDB指出的可以避免幻讀是怎么回事呢？

http://dev.mysql.com/doc/refman/5.0/en/innodb-record-level-locks.html

By default, InnoDB operates in REPEATABLE READ transaction isolation level and with the innodb_locks_unsafe_for_binlog system variable disabled. In this case, InnoDB uses next-key locks for searches and index scans, which prevents phantom rows (see Section 13.6.8.5, “Avoiding the Phantom Problem Using Next-Key Locking”).

准備的理解是，當隔離級別是可重復讀，且禁用innodb_locks_unsafe_for_binlog的情況下，在搜索和掃描index的時候使用的next-key locks可以避免幻讀。

關鍵點在於，是InnoDB默認對一個普通的查詢也會加next-key locks，還是說需要應用自己來加鎖呢？如果單看這一句，可能會以為InnoDB對普通的查詢也加了鎖，如果是，那和序列化（SERIALIZABLE）的區別又在哪里呢？

MySQL manual里還有一段：

13.2.8.5. Avoiding the Phantom Problem Using Next-Key Locking (http://dev.mysql.com/doc/refman/5.0/en/innodb-next-key-locking.html)

To prevent phantoms, InnoDB uses an algorithm called next-key locking that combines index-row locking with gap locking.

You can use next-key locking to implement a uniqueness check in your application: If you read your data in share mode and do not see a duplicate for a row you are going to insert, then you can safely insert your row and know that the next-key lock set on the successor of your row during the read prevents anyone meanwhile inserting a duplicate for your row. Thus, the next-key locking enables you to “lock” the nonexistence of something in your table.

我的理解是說，InnoDB提供了next-key locks，但需要應用程序自己去加鎖。manual里提供一個例子：

SELECT * FROM child WHERE id > 100 FOR UPDATE;

這樣，InnoDB會給id大於100的行（假如child表里有一行id為102），以及100-102，102+的gap都加上鎖。

可以使用show innodb status來查看是否給表加上了鎖。

 

下面看列子

 

例子1

a	b
SET SESSION TRANSACTION ISOLATION LEVEL REPEATABLE READ;
SET AUTOCOMMIT=0;
BEGIN	BEGIN
SELECT * FROM test WHERE a='1' FOR UPDATE;
	SELECT * FROM test
	INSERT test VALUES(1,1);
	鎖住了
INSERT test VALUES(1,1);
成功
COMMIT
	COMMIT
	避免幻讀可以select鎖住，再insert

 

 

 例子2

 

a	b
SET SESSION TRANSACTION ISOLATION LEVEL REPEATABLE READ;
SET AUTOCOMMIT=0;
BEGIN	BEGIN
SELECT * FROM test WHERE a='1' FOR UPDATE;
	SELECT * FROM test
	INSERT test VALUES(2,2);
	連2也被鎖住了？
INSERT test VALUES(1,1);
成功
COMMIT
	這次提交成功
	COMMIT
其他嘗試，這種情況無論插入2還是5都被鎖住等等

 

 例子3

 

a	b
SET SESSION TRANSACTION ISOLATION LEVEL REPEATABLE READ;
SET AUTOCOMMIT=0;
BEGIN	BEGIN
SELECT * FROM test	SELECT * FROM test
SELECT * FROM test WHERE a='1' FOR UPDATE;
	SELECT * FROM test
	INSERT test VALUES(2,2);
COMMIT	COMMIT
成功
COMMIT
	COMMIT

 

 

 例子 4

 

 

a	b
SET SESSION TRANSACTION ISOLATION LEVEL REPEATABLE READ;
SET AUTOCOMMIT=0;
BEGIN	BEGIN
SELECT * FROM test	SELECT * FROM test
SELECT * FROM test WHERE a='2' FOR UPDATE;
	SELECT * FROM test
	INSERT test VALUES(2,2);
	INSERT test VALUES(5,5);
COMMIT	COMMIT

 

例子 5

 

 

a	b
SET SESSION TRANSACTION ISOLATION LEVEL REPEATABLE READ;
SET AUTOCOMMIT=0;
BEGIN	BEGIN
SELECT * FROM test	SELECT * FROM test
SELECT * FROM test WHERE a='1' FOR UPDATE;
	INSERT test VALUES(5,5);
	插入5成功了
	UPDATE test SET b=33 WHERE a='3'
	INSERT test VALUES(2,2);
	2也可以
	UPDATE test SET b=11 WHERE a='1'
	1鎖住了
COMMIT
	COMMIT
SELECT * FROM test	SELECT * FROM test

 

 

 

以上例子說明，forupdate時候，id為主鍵，RR策略時候，鎖住了的條件符合的行，但是如果條件找不到任何列，鎖住的是整個表，（主鍵，唯一索引，非唯一索引，（insert，update對於gab鎖不通），參考第一章，第七章,第九章）

 

 

 ------------------------------------------------------------------

  再來看大神的解釋 ：鏈接: http://blog.bitfly.cn/post/mysql-innodb-phantom-read/

 

再看一個實驗，要注意，表t_bitfly里的id為主鍵字段。實驗三：

t Session A                 Session B
|
| START TRANSACTION;        START TRANSACTION;
|
| SELECT * FROM t_bitfly
| WHERE id<=1
| FOR UPDATE;
| +------+-------+
| | id   | value |
| +------+-------+
| |    1 | a     |
| +------+-------+
|                           INSERT INTO t_bitfly
|                           VALUES (2, 'b');
|                           Query OK, 1 row affected
|
| SELECT * FROM t_bitfly;
| +------+-------+
| | id   | value |
| +------+-------+
| |    1 | a     |
| +------+-------+
|                           INSERT INTO t_bitfly
|                           VALUES (0, '0');
|                           (waiting for lock ...
|                           then timeout)
|                           ERROR 1205 (HY000):
|                           Lock wait timeout exceeded;
|                           try restarting transaction
|
| SELECT * FROM t_bitfly;
| +------+-------+
| | id   | value |
| +------+-------+
| |    1 | a     |
| +------+-------+
|                           COMMIT;
|
| SELECT * FROM t_bitfly;
| +------+-------+
| | id   | value |
| +------+-------+
| |    1 | a     |
| +------+-------+
v

可以看到，用id<=1加的鎖，只鎖住了id<=1的范圍，可以成功添加id為2的記錄，添加id為0的記錄時就會等待鎖的釋放。

MySQL manual里對可重復讀里的鎖的詳細解釋：

http://dev.mysql.com/doc/refman/5.0/en/set-transaction.html#isolevel_repeatable-read

For locking reads (SELECT with FOR UPDATE or LOCK IN SHARE MODE),UPDATE, and DELETE statements, locking depends on whether the statement uses a unique index with a unique search condition, or a range-type search condition. For a unique index with a unique search condition, InnoDB locks only the index record found, not the gap before it. For other search conditions, InnoDB locks the index range scanned, using gap locks or next-key (gap plus index-record) locks to block insertions by other sessions into the gaps covered by the range.

------

一致性讀和提交讀，先看實驗，實驗四：

t Session A                      Session B
|
| START TRANSACTION;             START TRANSACTION;
|
| SELECT * FROM t_bitfly;
| +----+-------+
| | id | value |
| +----+-------+
| |  1 | a     |
| +----+-------+
|                                INSERT INTO t_bitfly
|                                VALUES (2, 'b');
|                                COMMIT;
|
| SELECT * FROM t_bitfly;
| +----+-------+
| | id | value |
| +----+-------+
| |  1 | a     |
| +----+-------+
|
| SELECT * FROM t_bitfly LOCK IN SHARE MODE;
| +----+-------+
| | id | value |
| +----+-------+
| |  1 | a     |
| |  2 | b     |
| +----+-------+
|
| SELECT * FROM t_bitfly FOR UPDATE;
| +----+-------+
| | id | value |
| +----+-------+
| |  1 | a     |
| |  2 | b     |
| +----+-------+
|
| SELECT * FROM t_bitfly;
| +----+-------+
| | id | value |
| +----+-------+
| |  1 | a     |
| +----+-------+
v

 

如果使用普通的讀，會得到一致性的結果，如果使用了加鎖的讀，就會讀到“最新的”“提交”讀的結果。

本身，可重復讀和提交讀是矛盾的。在同一個事務里，如果保證了可重復讀，就會看不到其他事務的提交，違背了提交讀；如果保證了提交讀，就會導致前后兩次讀到的結果不一致，違背了可重復讀。

可以這么講，InnoDB提供了這樣的機制，在默認的可重復讀的隔離級別里，可以使用加鎖讀去查詢最新的數據。

http://dev.mysql.com/doc/refman/5.0/en/innodb-consistent-read.html

If you want to see the “freshest” state of the database, you should use either the READ COMMITTED isolation level or a locking read:
SELECT * FROM t_bitfly LOCK IN SHARE MODE;

 

 

結論：MySQL InnoDB的可重復讀並不保證避免幻讀，需要應用使用加鎖讀來保證。而這個加鎖度使用到的機制就是next-key locks。

 

結論：mysql 的重復讀解決了幻讀的現象，但是需要 加上 select for update/lock in share mode 變成當前讀避免幻讀，普通讀select存在幻讀