Configure auto-recovery and reinstall nodes in a Managed Kubernetes cluster

When a node is reinstalled, its operating system is reinstalled, and all data on the boot disk will be deleted. Reinstalling all nodes in a group can be performed automatically via node auto-recovery—immediately for all nodes in a cluster. You can reinstall one or several nodes manually.

Configure node auto-recovery

For your information

Auto-recovery is unavailable:

in a Managed Kubernetes cluster with a single worker node;
in a node group on dedicated servers.

carefully

When a node undergoes auto-recovery, all data on the boot disk will be deleted.

When the auto-recovery option is enabled, worker nodes will be automatically reinstalled if they do not respond to health checks for approximately 15 minutes. During the recovery process, pods stop running on the unresponsive node and are redistributed to other nodes. Auto-recovery works for all nodes in the cluster simultaneously.

Node auto-recovery can be enabled when creating a cluster or for an existing cluster.

In the Control panel, from the top menu, click Products and select Managed Kubernetes.
Open the cluster page → Settings tab.
In the Automation block, enable Node auto-recovery.

Reinstall nodes manually

carefully

When a node is reinstalled, all data on the boot disk will be deleted.

You can reinstall each node manually.

In the Control panel, from the top menu, click Products and select Managed Kubernetes.
Open the cluster page → Cluster Nodes tab.
In the node menu, select Reinstall.