Enable cluster and machine maintenance mode

Before performing node maintenance operations that are not managed by MOSK, such as operating system configuration or node reboot, enable maintenance mode on the cluster and required machines using the MOSK management console or CLI to prepare workloads for maintenance.

Note

Maintenance mode that you enable manually is not the only way to put cluster workloads into maintenance. Operations managed by MOSK, such as graceful reboot and cluster update, also create a ClusterMaintenanceRequest object for the duration of the operation. As a result, the cluster services switch to the maintenance state, as if you enabled maintenance mode, even though the maintenance parameter is not set to true in the Cluster object.

Therefore, when a procedure has a separate path for the cluster in maintenance, such as Disable a machine, follow that path not only when maintenance is enabled explicitly but also while a graceful reboot or cluster update is in progress. To verify whether maintenance requests exist:

kubectl --kubeconfig <mosk-cluster-kubeconfig> get clustermaintenancerequests

• Enable maintenance mode on a cluster and machine using management console
• Enable maintenance mode on a cluster and machine using CLI