NCA-GENL cheat sheet

Nutanix Enterprise Cloud combines AOS (distributed storage and cluster services), AHV (native hypervisor), and Prism (management plane) into one software stack running on commodity x86 nodes.

Cluster scaling is scale-out: add a node to increase CPU, memory, and storage capacity simultaneously; there is no separate SAN or storage controller to expand.

A Nutanix cluster requires a minimum of three nodes for production resiliency; single-node clusters exist only for edge/ROBO use cases with reduced redundancy.

The Distributed Storage Fabric (DSF) pools local disks from every node into a single logical datastore presented to all hypervisors in the cluster.

Nutanix supports AHV, ESXi, and Hyper-V as hypervisors; AOS storage services are consumed by ESXi/Hyper-V via NFS/SMB datastores rather than a native controller integration.

Licensing tiers (Starter, Pro, Ultimate) govern feature access such as data-at-rest encryption, async replication, and Prism Central multi-cluster management.

Prism Element is per-cluster; Prism Central is a separate VM appliance that manages many clusters and enables categories, policies, and self-service.

Nutanix Cloud Clusters (NC2) extends the same AOS/AHV stack onto public cloud infrastructure such as AWS and Azure for burst and migration use cases.

Each node runs a Controller VM (CVM) that owns the local disks and serves I/O to the hypervisor over the internal network; the hypervisor host does not directly manage storage.

The CVM and the hypervisor host communicate over a private internal bridge; storage traffic between CVMs uses the backplane or the 10/25 GbE data network.

AOS metadata (extent map, etc.) is distributed and replicated so any CVM can locate any block without a central lookup service bottleneck.

Data is split into extents and stored as write-once immutable files on the Distributed Storage Fabric; overwrites create new extents and free old ones.

Replication Factor (RF) defaults to 2 or 3; RF3 tolerates two simultaneous disk or node failures, RF2 tolerates one.

The OpLog is a per-CVM persistent write buffer that absorbs random writes and coalesces them into sequential extent writes, reducing read-modify-write penalty.

Cassandra stores cluster metadata and configuration; Zookeeper handles cluster membership, leader election, and distributed locking.

Genesis is the bootstrap service that forms a cluster from individual nodes during foundation/imaging.

Stargate is the data I/O service on each CVM handling read/write requests from the hypervisor and remote CVMs.

Prism Element provides single-cluster health, alerts, VM management, storage capacity, and performance charts; Prism Central aggregates multiple clusters.

Prism Central features include Categories, Projects, Self-Service (Calm), Nutanix Guest Tools management, and policy-based protection domains across clusters.

Health checks in Prism run scheduled NCC (Nutanix Cluster Check) scans; failures surface as alerts with remediation guidance and knowledge-base links.

Alerts have severity levels (Critical, Warning, Info); they can be acknowledged, resolved, or routed to SNMP, syslog, or email targets.

Prism Central can be deployed as a single VM or a three-VM scale-out cluster for production resiliency; small deployments may use a single VM.

Role-Based Access Control (RBAC) in Prism assigns roles (Viewer, Cluster Admin, Super Admin, etc.) to users or AD/LDAP groups scoped by category.

The Tasks view records long-running operations (VM clone, snapshot, replication); completed tasks are retained per configured retention policy.

Prism APIs (v2 and v3) expose REST endpoints for automation; v3 is the current generation with intent-based specifications.

Containers are logical storage pools presented as NFS/SMB datastores; each container has its own RF, compression, deduplication, and encryption settings.

Inline compression and deduplication run in the CVM data path before data is written to disk; capacity savings are visible per container in Prism.

Erasure Coding (EC-X) replaces RF replication with parity stripes for cold data; it improves capacity efficiency but increases rebuild cost and is not recommended for latency-sensitive workloads.

Thin provisioning is default; a container's advertised size can exceed physical capacity because space is allocated on write.

Storage QoS policies limit IOPS per VM or per container to prevent noisy-neighbor effects on shared clusters.

Nutanix Volumes presents block storage (iSCSI) to external hosts that are not part of the AHV/ESXi cluster, useful for bare-metal or legacy workloads.

Nutanix Files provides a scale-out SMB/NFS file service delivered by separate File Server VMs on the AOS cluster.

Data-at-rest encryption (software or SED-based) is configured per cluster and requires the encryption capability in the license tier.

Local snapshots are metadata-light point-in-time copies stored on the same cluster; they protect against logical errors but not against site loss.

Protection Domains group VMs and consistency groups; a schedule defines snapshot frequency, retention, and optional remote replication.

Async replication sends snapshots to a remote cluster over the network at a configured RPO; sync replication is available only in specific metro configurations.

NearSync replication provides RPO measured in seconds (typically 1-15 minutes) and requires low-latency links between sites.

Nutanix Leap provides orchestration for disaster recovery: runbook automation, one-click failover/failback, and non-disruptive DR testing.

Mine is the integrated backup service that copies snapshots to object storage targets (S3-compatible, NFS) for long-term retention.

Snapshot retention is defined by a retention policy (e.g., keep hourly for N, daily for N); older snapshots are pruned automatically.

Metro Availability stretches a cluster across two sites with synchronous replication, giving zero RPO but requiring low-latency links and a witness for quorum.

AHV uses Open vSwitch (OVS) as the virtual switch; each host has a bridge (br0) that carries management and VM traffic.

A Nutanix VLAN is a logical network backed by a VLAN ID on the physical switch; VMs attach to VLAN-backed networks via virtual NICs.

The default AHV network is the 'vlan.0' untagged network; additional VLANs must exist on the upstream switch with matching IDs.

AHV supports bond modes including active-backup, balance-slb, and LACP (802.3ad); LACP requires switch-side LAG configuration.

Live migration (vMotion equivalent) in AHV is called 'live migration' and moves a running VM between hosts without downtime, subject to CPU compatibility.

AHV supports GPU passthrough and vGPU; the host must have the GPU and the correct driver in the guest.

Nutanix Guest Tools (NGT) installs drivers and agents inside the guest for VSS snapshots, self-service restore, and file-level restore.

AHV virtual machines can be protected by affinity/anti-affinity policies to control placement across hosts for availability.

Nutanix Calm provides application lifecycle management: blueprints define multi-tier apps and their deployment, scaling, and deletion actions.

Calm blueprints support Linux/Windows, AHV/ESXi, and public cloud providers, enabling hybrid deployments from a single definition.

Nutanix Xi Cloud Services (Xi Leap, Xi Epoch, etc.) are SaaS offerings; Xi Leap provides DR to a Nutanix-operated cloud.

NC2 (Nutanix Cloud Clusters) runs AOS/AHV natively on AWS or Azure, allowing live migration of VMs between on-prem and cloud.

Nutanix Move migrates VMs from ESXi, Hyper-V, or AWS/Azure into AHV or NC2; it supports online migration with minimal downtime.

Prism Central's self-service (Calm) uses Projects and Categories to scope which users can deploy which blueprints to which clusters.

REST API automation uses the v3 API; the Nutanix CLI (ncli) and acli provide command-line administration for legacy and current workflows.

Nutanix Marketplace and Calm allow publishing blueprints for reuse across projects and teams.

Nutanix Support Portal (portal.nutanix.com) provides case management, knowledge base, and firmware compatibility matrices.

NCC (Nutanix Cluster Check) is a health-check framework; run it before upgrades and after configuration changes to catch known issues.

LCM (Life Cycle Manager) in Prism Central manages AOS, AHV, firmware, and NCC updates across clusters with pre-check and rolling upgrade.

AOS upgrades are rolling: one CVM at a time is upgraded while the cluster remains available, provided RF and capacity headroom are sufficient.

Nutanix Cluster Health and Pulse send telemetry to Nutanix for proactive support; Pulse can be disabled but reduces support capability.

Security baseline: change default admin credentials, enable STIG/security configuration guides, and restrict Prism/API access to management networks.

Data-at-rest encryption uses either self-encrypting drives (SED) or software-based encryption; key management can be local or external (KMIP).

Nutanix recommends separating management, storage, and VM traffic onto dedicated VLANs and using 10/25 GbE links for production clusters.