This utility scans the process lists for the servers specified
using instances of the
--server option and selects
those that match the conditions specified using the
--match-xxx options. For a process to match,
all conditions given must match. The utility then either prints
the selected processes (the default) or executes certain actions
--match-xxx options are given, the utility
selects all processes.
--match-xxx options correspond to the
columns in the
table. For example,
--match-command specifies a
matching condition for
values. There is no
--match-time option. To
specify a condition based on process time, use
Processes that can be seen and killed are subject to whether the account used to connect to the server has the PROCESS and SUPER privileges. Without PROCESS, the account cannot see processes belonging to other accounts Without SUPER, the account cannot kill processes belonging to other accounts
To specify how to display output, use one of the following values
Display output in grid or table format like that of the mysql monitor.
Display output in comma-separated values format.
Display output in tab-separated format.
Display output in single-column format like that of the
\G command for the mysql
mysqlprocgrep accepts the following command-line options:
Select only processes that have been in the current state more
than a given time. The time value can be specified in two
formats: either using the
with hours and minutes optional, or as a sequence of numbers
with a suffix giving the period size.
The permitted suffixes are s (second), m (minute), h (hour), d (day), and w (week). For example, 4h15m mean 4 hours and 15 minutes.
For both formats, the specification can optionally be preceded
+ means older than the given time, and
- means younger than the given time.
Instead of displaying the selected processes, emit the
SELECT statement that
retrieves information about them. If the
--kill-query option is
given, the utility generates a stored procedure named
kill_processes() for killing the queries
rather than a SELECT
--sql, but produces
the output as the body of a stored procedure without the
CREATE PROCEDURE part of
the definition. This could be used, for example, to generate
an event for the server Event Manager.
When used with a kill option, code for killing the matching
queries is generated. Note that it is not possible to execute
the emitted code unless it is put in a stored routine, event,
or trigger. For example, the following code could be generated
to kill all idle connections for user
$ mysqlprocgrep --kill-connection --sql-body \ > --match-user=www-data --match-state=sleep DECLARE kill_done INT; DECLARE kill_cursor CURSOR FOR SELECT Id, User, Host, Db, Command, Time, State, Info FROM INFORMATION_SCHEMA.PROCESSLIST WHERE user LIKE 'www-data' AND State LIKE 'sleep' OPEN kill_cursor; BEGIN DECLARE id BIGINT; DECLARE EXIT HANDLER FOR NOT FOUND SET kill_done = 1; kill_loop: LOOP FETCH kill_cursor INTO id; KILL CONNECTION id; END LOOP kill_loop; END; CLOSE kill_cursor;
--format option, the
permitted values are not case sensitive. In addition, values may
be specified as any unambiguous prefix of a valid value. For
specifies the grid format. An error occurs if a prefix matches
more than one valid value.
The path to the MySQL client tools should be included in the PATH environment variable in order to use the authentication mechanism with login-paths. This will allow the utility to use the my_print_defaults tools which is required to read the login-path values from the login configuration file (.mylogin.cnf).
For each example, assume that the
root user on
localhost has sufficient privileges to kill
queries and connections.
Kill all queries created by user
mats that are
younger than 1 minute:
mysqlprocgrep --server=root@localhost \ --match-user=mats --age=-1m --kill-query
Kill all connections that have been idle for more than 1 hour:
mysqlprocgrep --server=root@localhost \ --match-command=sleep --age=1h --kill-connection