【Mysql】pt工具之-pt-kill

最新推荐文章于 2025-10-31 09:37:59 发布

转载最新推荐文章于 2025-10-31 09:37:59 发布 · 449 阅读

标签

#数据库 #shell

pt-kill

概述

从show processlist 中获取满足条件的连接或者从包含show processlist的文件中读取满足条件的连接并打印或者杀掉或者执行其他操作，这个工具在工作中实用性很高，当服务器连接出现异常后第一想到的就是pt-kill，自己写一个脚本也可以，但是功能还是没有pt-kill强大；

常用功能介绍

[root@hostnfsd :/root]$ pt-kill --defaults-file=/data/mydata/my_3307.cnf -uroot -p`cat /etc/sqlpass` --busy-time 60 --kill --victims all --interval 10 每隔10s 杀掉处于runnning状态超过60s的连接数；
[root@hostnfsd :/root]$ pt-kill --defaults-file=/data/mydata/my_3307.cnf -uroot -p`cat /etc/sqlpass` --match-command Sleep --kill --victims all --interval 10 每隔10s 杀掉处于sleep状态的连接数；
[root@hostnfsd :/root]$ pt-kill --defaults-file=/data/mydata/my_3307.cnf -uroot -p`cat /etc/sqlpass` --idle-time 60 --kill --victims all --interval 10 每隔10s 杀掉已经sleep达到60秒的连接数；

匹配状态stat处于sendind data的
[root@HaoDai_App_DB01 toolsqldir]# pt-kill --defaults-file=/data/mydata/my_3307.cnf -uroot -p`cat /etc/sqlpass` --match-state Sending --kill --victims all --interval 5

参数
–busy-time
运行时间
–idle-time
空闲时间
–victims
所有匹配的连接，对应有最久的连接
–interval
间隔时间，默认30s，有点长，可以根据实际情况来调节
–print
打印出来kill掉的连接
–match-command
匹配当前连接的命令
Query
Sleep
Binlog Dump
Connect
Delayed insert
Execute
Fetch
Init DB
Kill
Prepare
Processlist
Quit
Reset stmt
Table Dump

–match-state

匹配当前连接的状态

Locked
login
copy to tmp table
Copying to tmp table
Copying to tmp table on disk
Creating tmp table
executing
Reading from net
Sending data
Sorting for order
Sorting result
Table lock
Updating
–match-info
使用正则表达式匹配符合的sql
–match-db –match-user –match-host
见名知意
常用用法
杀掉空闲链接
pt-kill –match-command Sleep –idle-time 5 –host –port –interval –print –kill –victims all
杀掉运行时间超过5s的链接
pt-kill –match-command Query –busy-time 5 –host –port –interval –print –kill –victims all
杀掉匹配某个规则的正在运行的sql
pt-kill –match-command Query –busy-time 5 –host –port –interval –print –kill –victims all –match-info
杀掉正在进行filesort的sql
pt-kill –match-command Query –match-state “Sorting result” busy-time 5 –host –port –interval –print –kill –victims all
杀掉正在Copying to tmp table的sql
pt-kill –match-command Query –match-state “Copying to tmp table” busy-time 5 –host –port –interval –print –kill –victims all

几个重要参数

--daemonize  放在后台以守护进程的形式运行；
--interval  多久运行一次，单位可以是s,m,h，d等默认是s
--victims 默认是oldest,只杀最古老的查询。这是防止被查杀是不是真的长时间运行的查询，他们只是长期等待。这种种匹配按时间查询，杀死一个时间最高值。 all 杀掉所有满足的线程 杀死所有，但最长的保留不杀 action
--kill                 杀掉连接并且退出
--kill-query           只杀掉连接执行的语句，但是线程不会被终止
--print                打印满足条件的语句

QUERY MATCHES

--busy-time 批次查询已运行的时间超过这个时间的线程； --idle-time 杀掉sleep 了多少时间的连接线程，必须在--match-command sleep时才有效

opition查看参数吧
Usage: pt-kill [OPTIONS] [DSN] Options: --ask-pass Prompt for a password when connecting to MySQL --charset=s -A Default character set --config=A Read this comma-separated list of config files; if specified, this must be the first option on the command line --create-log-table Create the --log-dsn table if it does not exist --daemonize Fork to the background and detach from the shell --database=s -D The database to use for the connection --defaults-file=s -F Only read mysql options from the given file --filter=s Discard events for which this Perl code doesn't return true --group-by=s Apply matches to each class of queries grouped by this SHOW PROCESSLIST column --help Show help and exit --host=s -h Connect to host (default localhost) --interval=m How often to check for queries to kill. Optional suffix s=seconds, m=minutes, h=hours, d=days; if no suffix, s is used. --log=s Print all output to this file when daemonized --log-dsn=d Store each query killed in this DSN --password=s -p Password to use when connecting --pid=s Create the given PID file --port=i -P Port number to use for connection --query-id Prints an ID of the query that was just killed --rds Denotes the instance in question is on Amazon RDS --run-time=m How long to run before exiting. Optional suffix s= seconds, m=minutes, h=hours, d=days; if no suffix, s is used. --sentinel=s Exit if this file exists (default /tmp/pt-kill- sentinel) --set-vars=A Set the MySQL variables in this comma-separated list of variable=value pairs --socket=s -S Socket file to use for connection --stop Stop running instances by creating the --sentinel file --[no]strip-comments Remove SQL comments from queries in the Info column of the PROCESSLIST (default yes) --user=s -u User for login if not current user --verbose -v Print information to STDOUT about what is being done --version Show version and exit --[no]version-check Check for the latest version of Percona Toolkit, MySQL, and other programs (default yes) --victims=s Which of the matching queries in each class will be killed (default oldest) --wait-after-kill=m Wait after killing a query, before looking for more to kill. Optional suffix s=seconds, m=minutes, h=hours, d=days; if no suffix, s is used. --wait-before-kill=m Wait before killing a query. Optional suffix s= seconds, m=minutes, h=hours, d=days; if no suffix, s is used. Actions: --execute-command=s Execute this command when a query matches --kill Kill the connection for matching queries --kill-query Kill matching queries --print Print a KILL statement for matching queries; does not actually kill queries Class Matches: --any-busy-time=m Match query class if any query has been running for longer than this time. "Longer than" means that if you specify 10, for example, the class will only match if there's at least one query that has been running for greater than 10 seconds. Optional suffix s=seconds, m= minutes, h=hours, d=days; if no suffix, s is used. --each-busy-time=m Match query class if each query has been running for longer than this time. "Longer than" means that if you specify 10, for example, the class will only match if each and every query has been running for greater than 10 seconds. Optional suffix s=seconds, m=minutes, h= hours, d=days; if no suffix, s is used. --query-count=i Match query class if it has at least this many queries Query Matches: --busy-time=m Match queries that have been running for longer than this time. Optional suffix s=seconds, m=minutes, h= hours, d=days; if no suffix, s is used. --idle-time=m Match queries that have been idle/sleeping for longer than this time. Optional suffix s=seconds, m=minutes, h=hours, d=days; if no suffix, s is used. --ignore-command=s Ignore queries whose Command matches this Perl regex --ignore-db=s Ignore queries whose db (database) matches this Perl regex --ignore-host=s Ignore queries whose Host matches this Perl regex --ignore-info=s Ignore queries whose Info (query) matches this Perl regex --[no]ignore-self Don't kill pt-kill's own connection (default yes) --ignore-state=s Ignore queries whose State matches this Perl regex ( default Locked) --ignore-user=s Ignore queries whose user matches this Perl regex --match-all Match all queries that are not ignored --match-command=s Match only queries whose Command matches this Perl regex --match-db=s Match only queries whose db (database) matches this Perl regex --match-host=s Match only queries whose Host matches this Perl regex --match-info=s Match only queries whose Info (query) matches this Perl regex --match-state=s Match only queries whose State matches this Perl regex --match-user=s Match only queries whose User matches this Perl regex --replication-threads Allow matching and killing replication threads --test-matching=a Files with processlist snapshots to test matching options against

来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/29096438/viewspace-2122904/，如需转载，请注明出处，否则将追究法律责任。

转载于:http://blog.itpub.net/29096438/viewspace-2122904/