rpk cluster maintenance

Interact with cluster maintenance mode.

Maintenance mode is a state that a node may be placed into in which the node may be shutdown or restarted with minimal disruption to client workloads. The primary use of maintenance mode is to perform a rolling upgrade in which each node is placed into maintenance mode prior to upgrading the node.

Use the enable and disable subcommands to place a node into maintenance mode or remove it, respectively. Only one node at a time may be in maintenance mode.

When a node is placed into maintenance mode the following occurs:

Leadership draining. All raft leadership is transferred to another eligible node, and the node in maintenance mode rejects new leadership requests. By transferring leadership off of the node in maintenance mode all client traffic and requests are directed to other nodes minimizing disruption to client workloads when the node is shutdown.

Currently leadership is not transferred for partitions with one replica.


rpk cluster maintenance [flags] [command]


Value Type Description

-h, --help


Help for maintenance.



Redpanda or rpk config file; default search paths are ~/.config/rpk/rpk.yaml, $PWD, and /etc/redpanda/redpanda.yaml.

-X, --config-opt


Override rpk configuration settings; '-X help' for detail or '-X list' for terser detail.



rpk profile to use.

-v, --verbose


Enable verbose logging.