参考了这篇文章的一些内容:
http://xm-king.iteye.com/blog/770721
记住以下这张表:
我在springdemo库里面建了一个表:
CREATE TABLE `tx` (
`id` bigint(11) NOT NULL auto_increment,
`num` bigint(11) default 0 COMMENT '用户名',
PRIMARY KEY (`id`)
) ENGINE=InnoDB AUTO_INCREMENT=1 DEFAULT CHARSET=utf8 COMMENT='事务隔离级别测试表';
Mysql通过以下语句可以查询
SELECT @@tx_isolation;
SELECT @@global.tx_isolation;
SELECT @@session.tx_isolation;
默认的是 REPEATABLE-READ
可以通过以下方式修改隔离级别:
set tx_isolation='read-committed';
mysql> select @@tx_isolation;
+-----------------+
| @@tx_isolation |
+-----------------+
| REPEATABLE-READ |
+-----------------+
1 row in set (0.00 sec)
mysql> set tx_isolation='read-committed';
Query OK, 0 rows affected (0.00 sec)
mysql> select @@tx_isolation;
+----------------+
| @@tx_isolation |
+----------------+
| READ-COMMITTED |
+----------------+
1 row in set (0.00 sec)
但是,注意以上的改动,只能针对当前会话。
另起一个客户端,输入
select @@tx_isolation;
+-----------------+
| @@tx_isolation |
+-----------------+
| REPEATABLE-READ |
+-----------------+
————————————————————————————————
以下是各种情况的测试:
READ-UNCOMMITTED
update tx set num=11 where id=1;
ERROR 1665 (HY000): Cannot execute statement: impossible to write to binary log since BINLOG_FORMAT = STATEMENT and at least one table uses a storage engine limited to row-based logging. InnoDB is limited to row-logging when transaction isolation level is READ COMMITTED or READ UNCOMMITTED.
READ-COMMITTED
update tx set num=11 where id=1;
ERROR 1665 (HY000): Cannot execute statement: impossible to write to binary log since BINLOG_FORMAT = STATEMENT and at least one table uses a storage engine limited to row-based logging. InnoDB is limited to row-logging when transaction isolation level is READ COMMITTED or READ UNCOMMITTED.
REPEATABLE-READ
A:
mysql> start transaction;
Query OK, 0 rows affected (0.00 sec) mysql> select * from tx; (B更新前)
+----+------+
| id | num |
+----+------+
| 1 | 1 |
| 2 | 2 |
| 3 | 3 |
+----+------+
3 rows in set (0.00 sec) mysql> select * from tx; (B更新后)
+----+------+
| id | num |
+----+------+
| 1 | 1 |
| 2 | 2 |
| 3 | 3 |
+----+------+
3 rows in set (0.00 sec) mysql> commit;
Query OK, 0 rows affected (0.00 sec)
mysql> select * from tx;
+----+------+
| id | num |
+----+------+
| 1 | 10 |
| 2 | 2 |
| 3 | 3 |
+----+------+
可以看出,A的事务中,看不到B对数据的更新。
同样是 REPEATABLE-READ
mysql> start transaction;
Query OK, 0 rows affected (0.00 sec) mysql> select * from tx;
+----+------+
| id | num |
+----+------+
| 1 | 10 |
| 2 | 2 |
| 3 | 3 |
+----+------+
3 rows in set (0.00 sec) mysql> select * from tx;
+----+------+
| id | num |
+----+------+
| 1 | 10 |
| 2 | 2 |
| 3 | 3 |
+----+------+
3 rows in set (0.00 sec)
mysql> rollback;
Query OK, 0 rows affected (0.00 sec)
mysql> select * from tx;
+----+------+
| id | num |
+----+------+
| 1 | 10 |
| 2 | 2 |
| 3 | 3 |
| 4 | 4 |
+----+------+
4 rows in set (0.00 sec)
另一个客户端,insert 了一条记录.
对于正常的repeatable-read级别,是有可能出现幻读的情况,也就是说,第二遍的时候,A能够读到新插入的数据。
但是,InnoDB和Falcon存储引擎通过多版本并发控制(MVCC,Multiversion Concurrency Control)机制解决了该问题。
所以我上面用InnoDB引擎创建的数据库,就没有出现幻读的情况。
SERIALIZABLE
这里注意,只需要对使用事务Transaction的客户端设置SERIALIZABLE ,其他客户端的级别是什么都行,比如REPEATABLE-READ .
A
mysql> select @@tx_isolation;
+----------------+
| @@tx_isolation |
+----------------+
| SERIALIZABLE |
+----------------+
1 row in set (0.00 sec) mysql> start transaction;
Query OK, 0 rows affected (0.00 sec) mysql> select * from tx;
+----+------+
| id | num |
+----+------+
| 1 | 10 |
| 2 | 2 |
| 3 | 3 |
| 4 | 4 |
| 5 | 5 |
+----+------+
5 rows in set (0.00 sec) mysql> rollback;
Query OK, 0 rows affected (0.00 sec)
B
mysql> select @@tx_isolation;
+-----------------+
| @@tx_isolation |
+-----------------+
| REPEATABLE-READ |
+-----------------+
1 row in set (0.00 sec) mysql> insert into tx values(5,5);
Query OK, 1 row affected (0.09 sec) mysql> insert into tx values(6,6); Query OK, 1 row affected (31.60 sec)
这里,有2点需要注意:
1. A没有使用(select)表tx的时候,B仍然能够向tx表中插入数据;
2. A使用(select)表tx之后,B再进行insert操作,就会hang住,直到A表transaction结束。所以可以看到B的insert操作耗时31秒。当然了,也可能超时失败。
比如:
mysql> insert into tx values(10,10);
ERROR 1205 (HY000): Lock wait timeout exceeded; try restarting transaction
以上这条语句,看起来是锁了整个表。但是其实深层次的目的,是为了保证事务的完整性,以及让A两次操作的REPEATABLE_READ;
比如 使用 select * from tx limit 1; 后面就仍然可以insert。因为不改变之前select的结果。
如果 select * from tx limit 10; 而最终只检出8条。这时候就不能insert,因为insert了的话,下次同样select得到的结果就不一样。
同理,如果select了某条记录,那么update同一条记录就不行,update其他的记录就可以。而且在事务中能够读到新更新的数据。
所以记住,事务的隔离级别的要求,只对事务过程中已经获取过的数据有关。跟没获取过、其他不可见的数据,无关。