In the life of a database administrator or backend developer, there comes a moment of panic: a runaway SELECT statement is locking a critical table, an UPDATE without a WHERE clause is rewriting millions of rows, or a stalled transaction is grinding the application to a halt. The solution? Terminate the offending process.

SHOW PROCESSLIST; Or for full, non-truncated queries:

KILL 12345; Or with the optional CONNECTION keyword (same effect):

| Command | Effect | | :--- | :--- | | KILL QUERY 12345; | Terminates the but keeps the connection open. | | KILL CONNECTION 12345; (or just KILL 12345 ) | Terminates the query and closes the connection. |

However, "killing" a MySQL query is not without risk. This piece covers how to do it, when it’s safe, and the potential consequences. Before you can kill a query, you need its Process ID (Thread ID) . Connect to MySQL via the command line or your admin interface and run:

SELECT CONCAT('KILL ', id, ';') FROM information_schema.processlist WHERE time > 300; Sometimes you run KILL but the query remains in the process list with a state like Killed or Checking table . This means MySQL has acknowledged the kill command but is waiting for an internal resource (e.g., disk I/O, row lock) to release.

SELECT CONCAT('KILL ', id, ';') AS kill_command FROM information_schema.processlist WHERE user = 'app_user' AND command != 'Sleep'; Copy the output and execute it.

Killing a MySQL query is a scalpel – precise and effective, but careless use can cut deep. Always pair it with logging and post-mortem analysis to understand why you needed to kill it in the first place.