This page describes how to apply Node Pool updates by queuing a reconfigure reboot. Follow this procedure when you need to roll out an updated Node Pool configuration, such as a new system OS image or GPU driver, to existing Nodes in a CoreWeave Kubernetes Service (CKS) cluster.
You can manually reboot Nodes in two ways:
- Standard Node reboot: A typical reboot to apply a system update or other change.
- Reconfigure reboot: A reboot to apply an updated Node Pool configuration to existing Nodes.
Some Node Pool modifications require you to both reconfigure and reboot the Nodes to apply updates, for example, a new system OS image or GPU driver update. See Manage Node Pool configuration for more details.
To apply these updates, you can queue a reconfigure reboot for the necessary Nodes. If you need to reboot Nodes manually to apply a system update or other change, without reconfiguring the Node, see Reboot Nodes.
A reconfigure reboot is the rollout step for several Node Pool changes. For the end-to-end procedure that surrounds it, see the following:
Prerequisites
Before you begin, ensure you have:
To queue a reconfigure reboot that waits for active workloads to stop, use the CoreWeave Intelligent CLI with the --safe and --reconfigure flags. Replace [NODE-NAMES] with a space-separated list of the Nodes you want to reboot:
The --safe flag queues the reboot until active workloads stop. CKS requests a cordon to prevent new workload scheduling while existing workloads finish. For more information about idle Nodes and pending reboots, see Safe versus immediate reboots.
The reconfiguration and reboot process can take more than an hour:
- During the reconfiguration and reboot, the Node
PhaseState condition moves to
production-reconfigure-powercycle-test, and the Node becomes unavailable.
- When the reconfiguration is complete, the Node
PhaseState condition returns to production and the Node is uncordoned.
Reboot Nodes in small batchesWhen you reboot Nodes, limit the number of Nodes you reboot at one time to avoid service interruptions. If you need to reboot 50 or more Nodes at a time, contact support for assistance. Optional flags
The following sections describe how to skip post-reboot validation or request an immediate reboot.
Skip post-reboot validation
To skip validation, add the --no-test flag to the reconfigure command. This uses the Node lifecycle state
production-reconfigure-powerreset.
By default, a reconfigure reboot includes post-reboot validation. A single validation cycle typically takes about 30 minutes, but retries can extend the total duration.
To request an immediate reconfigure reboot without a confirmation prompt, omit --safe and use --force:
An immediate reboot doesn’t wait for active workloads to stop and can interrupt them. Without --safe or --force, the command also requests an immediate reboot, but prompts you if any selected Nodes are active. For the prompt options, see Request a Node reboot. Last modified on October 2, 2026
