为什么要避免大事务以及大事务如何解决？

lyb-geek

发布于 2022-03-04 13:36:54

6750

发布于 2022-03-04 13:36:54

文章被收录于专栏：Linyb极客之路Linyb极客之路

加关注这种话银家怎么好意思说出口嘛

什么是大事务 运行时间比较长，长时间未提交的事务就可以称为大事务

大事务产生的原因

操作的数据比较多

大量的锁竞争

事务中有其他非DB的耗时操作

大事务造成的影响

并发情况下，数据库连接池容易被撑爆
锁定太多的数据，造成大量的阻塞和锁超时
执行时间长，容易造成主从延迟
回滚所需要的时间比较长
undo log膨胀

如何查询大事务

注：本文的sql的操作都是基于mysql5.7版本

以查询执行时间超过10秒的事务为例：

select * from information_schema.innodb_trx where TIME_TO_SEC(timediff(now(),trx_started))>10select * from information_schema.innodb_trx where TIME_TO_SEC(timediff(now(),trx_started))>10

如何避免大事务

1、通用解法

在一个事务里面，避免一次处理太多数据
在一个事务里面，尽量避免不必要的查询
在一个事务里面，避免耗时太多的操作，造成事务超时。一些非DB的操作，比如rpc调用，消息队列的操作尽量放到事务之外操作

2、基于mysql5.7的解法

在InnoDB事务中，行锁是在需要的时候才加上的，但并不是不需要了就立刻释放，而是要等到事务结束时才释放。 如果你的事务中需要锁多个行，要把最可能造成锁冲突、最可能影响并发度的锁尽量往后放
通过SETMAX_EXECUTION_TIME命令，来控制每个语句查询的最长时间，避免单个语句意外查询太长时间
监控 information_schema.Innodb_trx表，设置长事务阈值，超过就报警/或者kill
在业务功能测试阶段要求输出所有的general_log，分析日志行为提前发现问题
设置innodb_undo_tablespaces值，将undo log分离到独立的表空间。如果真的出现大事务导致回滚段过大，这样设置后清理起来更方便

附录查询事务相关语句

注：sql语句都是基于mysql5.7版本

# 查询所有正在运行的事务及运行时间
select t.*,to_seconds(now())-to_seconds(t.trx_started) idle_time from INFORMATION_SCHEMA.INNODB_TRX t

# 查询事务详细信息及执行的SQL
select now(),(UNIX_TIMESTAMP(now()) - UNIX_TIMESTAMP(a.trx_started)) diff_sec,b.id,b.user,b.host,b.db,d.SQL_TEXT from information_schema.innodb_trx a inner join information_schema.PROCESSLIST b
on a.TRX_MYSQL_THREAD_ID=b.id and b.command = 'Sleep'
inner join performance_schema.threads c ON b.id = c.PROCESSLIST_ID
inner join performance_schema.events_statements_current d ON d.THREAD_ID = c.THREAD_ID;

# 查询事务执行过的所有历史SQL记录
SELECT
  ps.id 'PROCESS ID',
  ps.USER,
  ps.HOST,
  esh.EVENT_ID,
  trx.trx_started,
  esh.event_name 'EVENT NAME',
  esh.sql_text 'SQL',
  ps.time 
FROM
  PERFORMANCE_SCHEMA.events_statements_history esh
  JOIN PERFORMANCE_SCHEMA.threads th ON esh.thread_id = th.thread_id
  JOIN information_schema.PROCESSLIST ps ON ps.id = th.processlist_id
  LEFT JOIN information_schema.innodb_trx trx ON trx.trx_mysql_thread_id = ps.id 
WHERE
  trx.trx_id IS NOT NULL 
  AND ps.USER != 'SYSTEM_USER' 
ORDER BY
  esh.EVENT_ID;
  
 # 简单查询事务锁
 select * from sys.innodb_lock_waits
 
 # 查询事务锁详细信息
 SELECT
  tmp.*,
  c.SQL_Text blocking_sql_text,
  p.HOST blocking_host 
FROM
  (
  SELECT
    r.trx_state wating_trx_state,
    r.trx_id waiting_trx_id,
    r.trx_mysql_thread_Id waiting_thread,
    r.trx_query waiting_query,
    b.trx_state blocking_trx_state,
    b.trx_id blocking_trx_id,
    b.trx_mysql_thread_id blocking_thread,
    b.trx_query blocking_query 
  FROM
    information_schema.innodb_lock_waits w
    INNER JOIN information_schema.innodb_trx b ON b.trx_id = w.blocking_trx_id
    INNER JOIN information_schema.innodb_trx r ON r.trx_id = w.requesting_trx_id 
  ) tmp,
  information_schema.PROCESSLIST p,
  PERFORMANCE_SCHEMA.events_statements_current c,
  PERFORMANCE_SCHEMA.threads t 
WHERE
  tmp.blocking_thread = p.id 
  AND t.thread_id = c.THREAD_ID 
  AND t.PROCESSLIST_ID = p.id # 查询所有正在运行的事务及运行时间select t.*,to_seconds(now())-to_seconds(t.trx_started) idle_time from INFORMATION_SCHEMA.INNODB_TRX t# 查询事务详细信息及执行的SQLselect now(),(UNIX_TIMESTAMP(now()) - UNIX_TIMESTAMP(a.trx_started)) diff_sec,b.id,b.user,b.host,b.db,d.SQL_TEXT from information_schema.innodb_trx a inner join information_schema.PROCESSLIST bon a.TRX_MYSQL_THREAD_ID=b.id and b.command = 'Sleep'inner join performance_schema.threads c ON b.id = c.PROCESSLIST_IDinner join performance_schema.events_statements_current d ON d.THREAD_ID = c.THREAD_ID;# 查询事务执行过的所有历史SQL记录SELECTps.id 'PROCESS ID',ps.USER,ps.HOST,esh.EVENT_ID,trx.trx_started,esh.event_name 'EVENT NAME',esh.sql_text 'SQL',ps.timeFROMPERFORMANCE_SCHEMA.events_statements_history eshJOIN PERFORMANCE_SCHEMA.threads th ON esh.thread_id = th.thread_idJOIN information_schema.PROCESSLIST ps ON ps.id = th.processlist_idLEFT JOIN information_schema.innodb_trx trx ON trx.trx_mysql_thread_id = ps.idWHEREtrx.trx_id IS NOT NULLAND ps.USER != 'SYSTEM_USER'ORDER BYesh.EVENT_ID;# 简单查询事务锁select * from sys.innodb_lock_waits# 查询事务锁详细信息SELECTtmp.*,c.SQL_Text blocking_sql_text,p.HOST blocking_hostFROM(SELECTr.trx_state wating_trx_state,r.trx_id waiting_trx_id,r.trx_mysql_thread_Id waiting_thread,r.trx_query waiting_query,b.trx_state blocking_trx_state,b.trx_id blocking_trx_id,b.trx_mysql_thread_id blocking_thread,b.trx_query blocking_queryFROMinformation_schema.innodb_lock_waits wINNER JOIN information_schema.innodb_trx b ON b.trx_id = w.blocking_trx_idINNER JOIN information_schema.innodb_trx r ON r.trx_id = w.requesting_trx_id) tmp,information_schema.PROCESSLIST p,PERFORMANCE_SCHEMA.events_statements_current c,PERFORMANCE_SCHEMA.threads tWHEREtmp.blocking_thread = p.idAND t.thread_id = c.THREAD_IDAND t.PROCESSLIST_ID = p.id

参考

https://juejin.im/post/5d837d2c6fb9a06b05182737

https://www.cnblogs.com/zhixiang-org-cn/p/12454275.html

本文参与?腾讯云自媒体同步曝光计划，分享自微信公众号。

原始发表：2020-07-19，如有侵权请联系?cloudcommunity@tencent.com 删除

sql