MySQL线程阻塞处理方法
MySQL线程阻塞处理方法
by 樂小生
出现的错误：
ERROR 1205 (HY000): Lock wait timeout exceeded; try restarting transaction
解决办法：
查看sleep的进程
mysql> show full processlist;
+-------+----------+-------------------+----------+---------+-------+----------+---------
| Id | User | Host | db | Command | Time | State | Info +-------+----------+-------------------+----------+---------+-------+----------+---------
| 32579 | official | 127.0.0.1:42690 | official | Sleep | 20926 | | NULL |
| 32715 | erp | 127.0.0.1:2078 | NULL | Sleep | 4570 | | NULL |
| 32716 | erp | 127.0.0.1:2079 | erp | Sleep | 3590 | | NULL |
| 32717 | erp | 127.0.0.1:2080 | erp | Sleep | 3577 | | NULL |
| 32732 | erp | 127.0.0.1:2081 | erp | Sleep | 3590 | | NULL |
| 32733 | erp | 127.0.0.1:2082 | NULL | Sleep | 1412 | | NULL |
| 32734 | erp | 127.0.0.1:2083 | erp | Sleep | 1160 | | NULL |
| 32735 | erp | 127.0.0.1:2084 | erp | Sleep | 985 | | NULL |
| 32736 | erp | 127.0.0.1:2085 | erp | Sleep | 1160 | | NULL |
| 32739 | erp | 127.0.0.1:2088 | erp | Sleep | 695 | | NULL |
| 32740 | erp | 127.0.0.1:2089 | erp | Sleep | 768 | | NULL |
| 32741 | erp | 127.0.0.1:2090 | erp | Sleep | 688 | | NULL |
| 32742 | erp | 127.0.0.1:2091 | erp | Sleep | 686 | | NULL |
| 32743 | erp | 127.0.0.1:2092 | erp | Sleep | 687 | | NULL |
| 32745 | erp | 127.0.0.1:2093 | erp | Sleep | 686 | | NULL |
| 32746 | erp | 127.0.0.1:2094 | erp | Sleep | 686 | | NULL |
| 32763 | root | localhost | erp | Query | 0 | starting | show full processlist |
| 32765 | erp | 127.0.0.1:58216 | erp | Sleep | 18 | | NULL |
| 32766 | erp | 127.0.0.1:58218 | erp | Sleep | 41 | | NULL |
| 32767 | erp | 127.0.0.1:58220 | erp | Sleep | 41 | | NULL |
| 32768 | erp | 127.0.0.1:58222 | erp | Sleep | 41 | | NULL |
| 32769 | erp | 127.0.0.1:58224 | erp | Sleep | 2 | | NULL |
| 32770 | erp | 127.0.0.1:2106 | NULL | Sleep | 153 | | NULL |
| 32771 | erp | 127.0.0.1:2107 | erp | Sleep | 148 | | NULL |
| 32772 | erp | 127.0.0.1:2108 | erp | Query | 18 | updating | UPDATE `erp_user` SET `pwd`='OyUHgt21gTP2/5uFgbKZtq==' WHERE (`id`='10113') |
没有看到正在执行的慢SQL记录线程，再去查看innodb的事务表INNODB_TRX，看下里面是否有正在锁定的事务线程，看看ID是否在show full processlist里面的sleep线程中，如果是，就证明这个sleep的线程事务一直没有commit 或者 rollback而是卡住了，我们需要手动kill掉。
mysql> SELECT * FROM information_schema.INNODB_TRX\G;
*************************** 1. row ***************************
trx_id: 20342
trx_state: RUNNING
trx_started: 2018-01-29 16:48:13
trx_requested_lock_id: NULL
trx_wait_started: NULL
trx_weight: 4
trx_mysql_thread_id: 32735
trx_query: NULL
trx_operation_state: NULL
trx_tables_in_use: 0
trx_tables_locked: 1
trx_lock_structs: 2
trx_lock_memory_bytes: 1136
trx_rows_locked: 1
trx_rows_modified: 2
trx_concurrency_tickets: 0
trx_isolation_level: REPEATABLE READ
trx_unique_checks: 1
trx_foreign_key_checks: 1
trx_last_foreign_key_error: NULL
trx_adaptive_hash_latched: 0
trx_adaptive_hash_timeout: 10000
trx_is_read_only: 0
trx_autocommit_non_locking: 0
1 row in set (0.00 sec)
ERROR:
No query specified
看到有这条 32735 的sql，kill掉，执行kill 32735;
kill 32735;
然后再去查询INNODB_TRX表，就没有阻塞的事务sleep线程存在了，如下所示：
mysql> SELECT * FROM information_schema.INNODB_TRX\G;
现在可以正常执行 sql 语句了
也可以更改INNODB_LOCK_WAIT_TIMEOUT的值，让其调大一点
select * from information_schema.global_variables where variable_name like '%timeout%';
+----------------------------+----------------+
| VARIABLE_NAME | VARIABLE_VALUE |
+----------------------------+----------------+
| DELAYED_INSERT_TIMEOUT | 300 |
| WAIT_TIMEOUT | 28800 |
| CONNECT_TIMEOUT | 10 |
| LOCK_WAIT_TIMEOUT | 31536000 |
| INNODB_ROLLBACK_ON_TIMEOUT | OFF |
| THREAD_POOL_IDLE_TIMEOUT | 60 |
| NET_WRITE_TIMEOUT | 60 |
| INNODB_LOCK_WAIT_TIMEOUT | 50 |
| INTERACTIVE_TIMEOUT | 28800 |
| DEADLOCK_TIMEOUT_LONG | 50000000 |
| SLAVE_NET_TIMEOUT | 3600 |
| DEADLOCK_TIMEOUT_SHORT | 10000 |
| NET_READ_TIMEOUT | 30 |
+----------------------------+----------------+
或者
MariaDB [(none)]> SHOW GLOBAL VARIABLES LIKE 'innodb_lock_wait_timeout';
+--------------------------+-------+
| Variable_name | Value |
+--------------------------+-------+
| innodb_lock_wait_timeout | 50 |
+--------------------------+-------+
数据库配置文件修改：
原因：原因是你使用的InnoDB 表类型的时候,
默认参数:innodb_lock_wait_timeout设置锁等待的时间是50s,
因为有的锁等待超过了这个时间,所以抱错.
你可以把这个时间加长,或者优化存储过程,事务避免过长时间的等待.
解决的办法有两个：
第一：innodb_lock_wait_timeout 锁定等待时间改大
my.cnf文件：
#innodb_lock_wait_timeout = 50
修改为
innodb_lock_wait_timeout = 500