MySQL 串行隔离级别小记

1. 串行加的是行级锁不是表锁

　　数据和隔离级别

mysql> select * from t2;
+---+
| a |
+---+
| 1 |
| 2 |
| 5 |
| 6 |
| 7 |
| 9 |
+---+
6 rows in set (5.09 sec)

mysql> show variables like '%isolation%';
+-----------------------+--------------+
| Variable_name         | Value        |
+-----------------------+--------------+
| transaction_isolation | SERIALIZABLE |
| tx_isolation          | SERIALIZABLE |
+-----------------------+--------------+
2 rows in set, 1 warning (0.00 sec)

　　事务A：

mysql> begin;
Query OK, 0 rows affected (0.00 sec)
mysql> select * from t2 where a = 2 for update;
+---+
| a |
+---+
| 2 |
+---+
1 row in set (0.00 sec)

　　事务B：

mysql> insert into t2 values(3);
Query OK, 1 row affected (0.01 sec)
mysql> insert into t2 values(2);
ERROR 1205 (HY000): Lock wait timeout exceeded; try restarting transaction

2. 当autocommit=1时，写是否阻塞读

　　事务A

mysql> begin;
Query OK, 0 rows affected (0.00 sec)
mysql> insert into t2 values(13);
Query OK, 1 row affected (0.00 sec)

　　事务B，不显式开启事务查询。查询的是事务A开启之前的快照

mysql> select *from t2;
+----+
| a  |
+----+
|  1 |
|  2 |
|  3 |
|  5 |
|  6 |
|  7 |
| 11 |
+----+
7 rows in set (0.00 sec)

　　事务B，显式开启事务查询

mysql> begin;
Query OK, 0 rows affected (0.00 sec)

mysql> select *from t2;
ERROR 1205 (HY000): Lock wait timeout exceeded; try restarting transaction

　　当autocommit=1时，不显式开启事务的读是快照读。

3. 当autocommit=1时，读是否会阻塞写

　　3.1 事务B，不显式开启事务查询：

#事务B
mysql> select *from t2;
+----+
| a  |
+----+
|  1 |
|  2 |
|  3 |
|  5 |
|  6 |
|  7 |
| 11 |
| 13 |
+----+
8 rows in set (0.00 sec)

#事务A
mysql> insert into t2 values(15);
Query OK, 1 row affected (0.02 sec)

　　3.2 事务B，显式开启事务查询：

#事务B
mysql> begin;
Query OK, 0 rows affected (0.00 sec)
mysql> select *from t2;
+----+
| a  |
+----+
|  1 |
|  2 |
|  3 |
|  5 |
|  6 |
|  7 |
| 11 |
| 13 |
| 15 |
+----+
9 rows in set (0.00 sec)

#事务A
mysql> insert into t2 values(16);
ERROR 1205 (HY000): Lock wait timeout exceeded; try restarting transaction

 　　串行隔离级别总结：

　　1. 写加的是记录锁。只要不是操作同一条记录，写写互不影响

　　2. 没有显式开启事务的读是快照，不会加锁

　　3. 显式开启事务的读会加共享锁，类似于select ... lock in share mode，此时读写互斥。

测试环境：mysql 5.7.26，引擎InnoDB

 参考：https://blog.csdn.net/cuiyandan4490/article/details/100464470