CLIENT KILL <ip:port | <[ID client-id] | [TYPE <NORMAL | MASTER | SLAVE | REPLICA | PUBSUB>] | [USER username] | [ADDR ip:port] | [LADDR ip:port] | [SKIPME <YES | NO>] [[ID client-id] | [TYPE <NORMAL | MASTER | SLAVE | REPLICA | PUBSUB>] | [USER username] | [ADDR ip:port] | [LADDR ip:port] | [SKIPME <YES | NO>] ...]>>
- Available since:
- Time complexity:
- O(N) where N is the number of client connections
- ACL categories:
CLIENT KILL command closes a given client connection. This command support two formats, the old format:
CLIENT KILL addr:port
ip:port should match a line returned by the
CLIENT LIST command (
The new format:
CLIENT KILL <filter> <value> ... ... <filter> <value>
With the new form it is possible to kill clients by different attributes instead of killing just by address. The following filters are available:
CLIENT KILL ADDR ip:port. This is exactly the same as the old three-arguments behavior.
CLIENT KILL LADDR ip:port. Kill all clients connected to specified local (bind) address.
CLIENT KILL ID client-id. Allows to kill a client by its unique
ID's are retrieved using the
CLIENT KILL TYPE type, where type is one of
pubsub. This closes the connections of all the clients in the specified class. Note that clients blocked into the
MONITORcommand are considered to belong to the
CLIENT KILL USER username. Closes all the connections that are authenticated with the specified ACL username, however it returns an error if the username does not map to an existing ACL user.
CLIENT KILL SKIPME yes/no. By default this option is set to
yes, that is, the client calling the command will not get killed, however setting this option to
nowill have the effect of also killing the client calling the command.
It is possible to provide multiple filters at the same time. The command will handle multiple filters via logical AND. For example:
CLIENT KILL addr 127.0.0.1:12345 type pubsub
is valid and will kill only a pubsub client with the specified address. This format containing multiple filters is rarely useful currently.
When the new form is used the command no longer returns
OK or an error, but instead the number of killed clients, that may be zero.
CLIENT KILL and Redis Sentinel
Recent versions of Redis Sentinel (Redis 2.8.12 or greater) use CLIENT KILL in order to kill clients when an instance is reconfigured, in order to force clients to perform the handshake with one Sentinel again and update its configuration.
Due to the single-threaded nature of Redis, it is not possible to kill a client connection while it is executing a command. From the client point of view, the connection can never be closed in the middle of the execution of a command. However, the client will notice the connection has been closed only when the next command is sent (and results in network error).
One of the following:
- Simple string reply:
OKwhen called in 3 argument format and the connection has been closed.
- Integer reply: when called in filter/value format, the number of clients killed.
- Starting with Redis version 2.8.12: Added new filter format.
- Starting with Redis version 2.8.12:
- Starting with Redis version 3.2.0: Added
mastertype in for
- Starting with Redis version 5.0.0: Replaced
slavestill supported for backward compatibility.
- Starting with Redis version 6.2.0: