How to find and kill misbehaving MySQL queries

Sometimes the complexity of a relational database system can be overwhelming. Fortunately, that complexity is an advantage, as with MySQL's tools for managing queries. In this tutorial, I will show you how to find and kill any misbehaving MySQL queries.

To view the currently-running queries, log in to the MySQL console and run the 'show processlist' command:

mysql> show processlist;
+--------+--------+-----------------+---------+---------+-------+-------+------------------+-----------+---------------+-----------+
| Id     | User   | Host            | db      | Command | Time  | State | Info             | Rows_sent | Rows_examined | Rows_read |
+--------+--------+-----------------+---------+---------+-------+-------+------------------+-----------+---------------+-----------+
|  78233 | root   | 127.0.0.1:37527 | mysql   | Sleep   | 16474 |       | NULL             |         6 |             6 |         6 |
|  84546 | root   | 127.0.0.1:48593 | mysql   | Sleep   | 13237 |       | NULL             |         2 |             2 |         2 |
| 107083 | root   | 127.0.0.1:56451 | mysql   | Sleep   | 15488 |       | NULL             |         1 |           121 |       121 |
| 131455 | root   | 127.0.0.1:48550 | NULL    | Query   |     0 | NULL  | show processlist |         0 |             0 |         0 |
+--------+--------+-----------------+---------+---------+-------+-------+------------------+-----------+---------------+-----------+
4 rows in set (0.03 sec)

The first column you should look at is 'Time', which is the number of seconds the process has been "doing the thing it's doing". A process whose command is 'Sleep' is waiting for a query to come in, so it's not consuming any resources. For any other process, however, a 'Time' of more than a few seconds indicates a problem.

In this case, the only query running is our 'show processlist' command. Let's see what it looks like if we have a poorly-written query running:

mysql> show processlist;
+--------+--------+-----------------+-----------+---------+-------+--------------+----------------------------------+-----------+---------------+-----------+
| Id     | User   | Host            | db        | Command | Time  | State        | Info                             | Rows_sent | Rows_examined | Rows_read |
+--------+--------+-----------------+-----------+---------+-------+--------------+----------------------------------+-----------+---------------+-----------+
|  78233 | root   | 127.0.0.1:37527 | example   | Sleep   | 18046 |              | NULL                             |         6 |             6 |         6 |
|  84546 | root   | 127.0.0.1:48593 | example   | Sleep   | 14809 |              | NULL                             |         2 |             2 |         2 |
| 107083 | root   | 127.0.0.1:56451 | example   | Sleep   | 17060 |              | NULL                             |         1 |           121 |       121 |
| 132033 | root   | 127.0.0.1:54642 | example   | Query   |    27 | Sending data | select max(subtotal) from orders |         0 |             0 |         0 |
| 133933 | root   | 127.0.0.1:48679 | NULL      | Query   |     0 | NULL         | show processlist                 |         0 |             0 |         0 |
| 134122 | root   | 127.0.0.1:49264 | example   | Sleep   |     0 |              | NULL                             |         0 |             0 |         0 |
+--------+--------+-----------------+-----------+---------+-------+--------------+----------------------------------+-----------+---------------+-----------+
6 rows in set (0.00 sec)

Ah! Now we see there is a query that's been running for almost 30 seconds. If we don't want to let it run its course, we can kill it by passing its 'Id' to the kill command:

mysql> kill 132033;
Query OK, 0 rows affected (0.00 sec)
mysql>

(Note that MySQL will always report 0 rows affected, because we're not altering any data.)

Judicious use of the kill command can clean up a backlog of queries. Remember, however, that it's not a permanent solution - if those queries came from your application, you need to rewrite them, or you'll continue to see the same issue reappear.

## See Also

MySQL's documentation on the different 'Command' values:
https://dev.mysql.com/doc/refman/5.7/en/thread-commands.html

Subscribe to Xmodulo

Do you want to receive Linux FAQs, detailed tutorials and tips published at Xmodulo? Enter your email address below, and we will deliver our Linux posts straight to your email box, for free. Delivery powered by Google Feedburner.

One thought on “How to find and kill misbehaving MySQL queries

Leave a comment

Your email address will not be published. Required fields are marked *