MySQL行級鎖測試

luashin發表於2016-01-08

    MySQL innodb儲存引擎使用與Oracle相同的行鎖機制,對如何檢視系統中存在的行鎖情況在下面的實驗中將會看到。下面是測試過程:
session 1:更新記錄

mysql> set autocommit=off;

Query OK, 0 rows affected (0.01 sec)

mysql> update t1 set email='test@test.com' where id=0;

Query OK, 4 rows affected (0.00 sec)

Rows matched: 4  Changed: 4  Warnings: 0


session 2:也更新相同的記錄,出現等待

 mysql> set autocommit=off;

 Query OK, 0 rows affected (0.00 sec)

 mysql> update t1 set email='abc' where id=0;

session 3:檢視系統等待事件:

mysql> show status like '%lock%';

+-------------------------------+---------+

| Variable_name                 | Value   |

+-------------------------------+---------+

| Com_lock_tables               | 0       |

| Com_unlock_tables             | 0       |

| Innodb_row_lock_current_waits | 1       | --這裡

| Innodb_row_lock_time          | 0       |

| Innodb_row_lock_time_avg      | 0       |

| Innodb_row_lock_time_max      | 0       |

| Innodb_row_lock_waits         | 1       |

| Key_blocks_not_flushed        | 0       |

| Key_blocks_unused             | 14497   |

| Key_blocks_used               | 0       |

| Qcache_free_blocks            | 1       |

| Qcache_total_blocks           | 1       |

| Table_locks_immediate         | 2070991 |

| Table_locks_waited            | 2       |

+-------------------------------+---------+

14 rows in set (0.01 sec)


session 1:提交記錄

mysql> commit;

Query OK, 0 rows affected (0.01 sec)


session 2:update立刻完成

mysql> update t1 set email='abc' where id=0;

Query OK, 4 rows affected (2 min 43.44 sec)  --這麼長時間完成更新操作

Rows matched: 4  Changed: 4  Warnings: 0

session 3:再次檢視系統等待事件

mysql> show status like '%lock%';

+-------------------------------+---------+

| Variable_name                 | Value   |

+-------------------------------+---------+

| Com_lock_tables               | 0       |

| Com_unlock_tables             | 0       |

| Innodb_row_lock_current_waits | 0       |  --這裡為0

| Innodb_row_lock_time          | 163436  |

| Innodb_row_lock_time_avg      | 163436  |

| Innodb_row_lock_time_max      | 163436  |

| Innodb_row_lock_waits         | 1       |

| Key_blocks_not_flushed        | 0       |

| Key_blocks_unused             | 14497   |

| Key_blocks_used               | 0       |

| Qcache_free_blocks            | 1       |

| Qcache_total_blocks           | 1       |

| Table_locks_immediate         | 2070991 |

| Table_locks_waited            | 2       |

+-------------------------------+---------+

14 rows in set (0.01 sec)


查詢會話session 1,session 2的連線ID

session 1:

mysql> status;

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

mysql  Ver 12.22 Distrib 4.0.24, for pc-solaris2.10 (i386)

 Connection id:          15

 session 2:

mysql> status;

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

mysql  Ver 12.22 Distrib 4.0.24, for pc-solaris2.10 (i386)

 Connection id:          13

在上面的session 1尚沒有提交的時候,可以執行下列命令,檢視一些事務阻塞資訊

mysql> show innodb status\G;

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

TRANSACTIONS

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

Trx id counter 0 3852351

Purge done for trx's n:o < 0 3852350 undo n:o < 0 0

History list length 11

Total number of lock structs in row lock hash table 7

LIST OF TRANSACTIONS FOR EACH SESSION:

---TRANSACTION 0 0, not started, OS thread id 15

MySQL thread id 18, query id 2071119 localhost root

show innodb status

---TRANSACTION 0 3852350, ACTIVE 6 sec, OS thread id 14 starting index read

mysql tables in use 1, locked 1

LOCK WAIT 2 lock struct(s), heap size 320

MySQL thread id 13, query id 2071118 localhost test Updating      --這裡可以看到等待者

update t1 set email='abc' where id=0          --這裡可以看到等待者正在執行的SQL

------- TRX HAS BEEN WAITING 6 SEC FOR THIS LOCK TO BE GRANTED:

RECORD LOCKS space id 0 page no 32782 n bits 1056 index `idx_t1_id` of table `dc_test/t1` trx id 0 3852350 lock_mode X waiting

Record lock, heap no 2 PHYSICAL RECORD: n_fields 2; compact format; info bits 0

 0: len 4; hex 80000000; asc     ;; 1: len 6; hex 00000196fe5d; asc      ];;

 

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

---TRANSACTION 0 3852348, ACTIVE 391 sec, OS thread id 12

7 lock struct(s), heap size 1024, undo log entries 4

MySQL thread id 15, query id 2071117 localhost test

來自 “ ITPUB部落格 ” ,連結:http://blog.itpub.net/9034054/viewspace-1973727/,如需轉載,請註明出處,否則將追究法律責任。

相關文章