Low-level routine maintenance such as applying a RHEL OS patch to the hosts in an NSP Kubernetes cluster may require that you stop and start Kubernetes in the cluster.

Stopping Kubernetes in an NSP cluster stops the NSP software in the cluster, and creates a network management outage in a standalone deployment.

In a DR deployment, you can avoid a network management outage by stopping and starting the clusters in sequence, as specified in Workflow: stop and start DR NSP clusters.

The following procedures describe the following basic Kubernetes administration operations for the NSP:

• starting and stopping clusters

• identifying the master node in a cluster

• displaying the DR status

• performing DR role switches

The NSP has internal agents that monitor the NSP deployment infrastructure. An alarm in raised in response to the detection of one of the following NSP deployer VM fault conditions:

Note: An NSP deployer VM unreachability alarm clears automatically when reachability is restored.

• NSP deployer VM unreachability

• Harbor registry unavailability

• excessive disk consumption