Nutanix NCP-CN flashcards
Nutanix NCP-CN · Build recall with a focused review session.
32 cards. Read the term and the definition here. The deck at the top of this page is optional.
- NKP management cluster
- The cluster that hosts NKP lifecycle and addon controllers. It is created first and is used to provision, upgrade, and delete workload clusters on the chosen infrastructure provider.
- Workload cluster
- A Kubernetes cluster created and managed by the NKP management cluster. It runs application workloads and receives addons and version updates from the management plane.
- Node pool
- A group of nodes with identical role and resource profile inside an NKP cluster. Scaling the pool adds or removes worker nodes without recreating the control plane.
- Etcd quorum rule
- Etcd requires a majority of members to be available for writes. With stacked etcd on control-plane nodes, keep the control-plane count odd to avoid split-brain and write outages.
- Declarative cluster definition
- A configuration object describing node pools, networking, and addons that is applied to the management cluster. The management plane reconciles actual infrastructure to match the declared desired state.
- Addon versioning
- CNI, CSI, load balancer, and policy addons are installed by the management plane and versioned with the NKP release, so upgrading NKP can change addon versions in workload clusters.
- CNI plugin role
- The CNI provides pod IP address management and pod-to-pod routing. Its implementation determines whether NetworkPolicy enforcement, BGP peering, or eBPF features are available.
- LoadBalancer address pool
- Service type LoadBalancer needs an implementation that assigns external addresses. On AHV this pool must come from a subnet and VLAN routable to clients.
- Service CIDR overlap
- Pod and Service CIDRs must not overlap with AHV subnets or peered networks. Overlap breaks routing and causes unreachable services after cluster creation.
- Headless Service
- A Service with clusterIP set to None. DNS returns pod IPs directly instead of a virtual IP, which supports StatefulSet peer discovery and database cluster formation.
- NetworkPolicy dependency
- NetworkPolicy objects only affect traffic if the deployed CNI enforces them. With a non-enforcing CNI, applying policies changes nothing on the data path.
- Overlay MTU mismatch
- Encapsulation headers reduce usable MTU. If the physical network MTU is not raised accordingly, large packets drop while small control packets succeed, causing intermittent stalls.
- Nutanix CSI driver
- A CSI-compatible driver that provisions persistent volumes on Nutanix Volumes or Nutanix Files. It runs a controller Deployment and a node DaemonSet using Prism credentials.
- StorageClass reclaimPolicy
- Determines what happens to the backing Nutanix disk when a PVC is deleted. Delete removes it; Retain preserves the volume for manual recovery or reuse.
- ReadWriteMany requirement
- Multiple nodes mounting the same PVC at once requires a shared filesystem backend such as Nutanix Files. Block-based Volumes only support ReadWriteOnce attachment.
- Volume expansion flag
- A StorageClass must set allowVolumeExpansion to true before a PVC can grow. The CSI driver then expands the Nutanix volume and filesystem online where supported.
- VolumeSnapshotClass
- Defines how CSI snapshots are created for a storage backend. Restoring a snapshot produces a new PVC, forming the basis of application data protection workflows.
- StatefulSet PVC retention
- StatefulSets create per-replica PVCs through volumeClaimTemplates. Deleting the StatefulSet leaves those PVCs and their Nutanix volumes in place unless removed explicitly.
- NDB software profile
- Defines the database engine and version binaries used when NDB provisions a database. It is paired with compute, network, and parameter profiles during provisioning.
- NDB Time Machine
- A policy that schedules snapshots and log capture with retention settings. It enables point-in-time recovery and provides the source snapshots for space-efficient database clones.
- NDB database clone
- A space-efficient copy created from a snapshot or Time Machine. Clones provision dev and test databases quickly without copying the full dataset.
- NDB compute profile
- Specifies vCPU and memory assigned to a provisioned database VM. It is selected alongside software, network, and parameter profiles during database creation.
- NDB role-based access
- NDB uses roles such as admin, operator, and user with entity-level permissions. Operations are recorded in audit logs to support compliance requirements.
- NC2 definition
- Nutanix Cloud Clusters runs the AHV and Prism stack on public cloud infrastructure as a managed cluster, delivering a consistent Nutanix experience in AWS or Azure.
- NC2 suspend versus delete
- Suspending an NC2 cluster reduces compute cost while retaining storage and configuration. Deleting removes the cluster and its cloud resources permanently.
- Hybrid CIDR planning
- On-premises and NC2 networks must use non-overlapping node, pod, and service CIDRs. Overlap is the most common blocker for hybrid connectivity and migration.
- Prism Central multi-cluster view
- Registering on-premises and NC2 clusters to one Prism Central gives a single management pane. Latency and CIDR planning must be validated before peering them.
- Nutanix Flow microsegmentation
- Flow enforces network policy at the AHV VM NIC level, including Kubernetes worker and control-plane VMs. It operates independently of Kubernetes NetworkPolicy objects.
- Two-layer Kubernetes backup
- Complete protection needs both Kubernetes API object backup and persistent volume data backup. A PVC snapshot alone does not capture namespaces, ConfigMaps, or other objects.
- Velero restore prerequisite
- Restoring backed-up Kubernetes objects requires matching StorageClasses, CRDs, and namespace configuration in the target cluster, or volumes and custom resources fail to restore.
- RPO versus RTO
- RPO is the maximum acceptable data loss, set by replication or snapshot frequency. RTO is the maximum acceptable downtime, set by restore or failover automation speed.
- Kubernetes Secrets encoding
- Secrets are base64-encoded, not encrypted, by default. Enable encryption at rest through an encryption provider and integrate an external secrets store where required.