【Mysql】pt工具之-pt-kill

pt-kill

概述

從show processlist 中獲取滿足條件的連線或者從包含show processlist的檔案中讀取滿足條件的連線並列印或者殺掉或者執行其他操作，這個工具在工作中實用性很高，當伺服器連線出現異常後第一想到的就是pt-kill，自己寫一個指令碼也可以，但是功能還是沒有pt-kill強大；

常用功能介紹

幾個重要引數

--daemonize  放在後臺以守護程式的形式執行；
--interval  多久執行一次，單位可以是s,m,h，d等預設是s
--victims 預設是oldest,只殺最古老的查詢。這是防止被查殺是不是真的長時間執行的查詢，他們只是長期等待。這種種匹配按時間查詢，殺死一個時間最高值。 all 殺掉所有滿足的執行緒 殺死所有，但最長的保留不殺 action
--kill                 殺掉連線並且退出
--kill-query           只殺掉連線執行的語句，但是執行緒不會被終止
--print                列印滿足條件的語句

	QUERY MATCHES

--busy-time 批次查詢已執行的時間超過這個時間的執行緒； --idle-time 殺掉sleep 了多少時間的連線執行緒，必須在--match-command sleep時才有效

	
	

		
		

			opition檢視引數吧 
		

		

			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