☁️ Nutanix Foundations Guide

Complete Nutanix HCI foundations — infrastructure, Prism, data resiliency, networking and operations

19 Sections HCI · AHV · AOS Prism Element & Central Foundation & Imaging

📘 Technical Learning Guide

Nutanix Foundations

Complete technical reference covering AOS, AHV, Prism Element, Prism Central, acli, ncli, Files, Flow, Leap, Calm, Foundation Imaging, and more.

AOS · AHV Prism Element Prism Central acli · ncli Nutanix Files Flow · Leap Self-Service Foundation DR · Protection

Sections

200+

Terms Defined

60+

CLI Commands

★

Lab Cheat Sheet

SECTION 01

Introduction & Evolution of Enterprise Infrastructure

1. Introduction & Evolution of Enterprise Infrastructure

1.1 Traditional Data Center Architecture

Before Nutanix, enterprise IT relied on a three-tier architecture comprising separate silos: storage, compute, and networking. Each tier was independently purchased, managed, and scaled — leading to significant complexity and inefficiency.

Three-Tier Model	Storage Array (SAN/NAS) + Blade/Rack Servers + Fibre Channel / Ethernet switches — all managed separately
Storage Array (SAN)	Centralized block storage connected via Fibre Channel or iSCSI. Examples: EMC VMAX, NetApp FAS, Pure Storage.
NAS	Network Attached Storage — file-level access via NFS/CIFS/SMB. Shared file systems for general workloads.
Hypervisor	VMware vSphere (ESXi), Microsoft Hyper-V, KVM — abstracts physical hardware into virtual machines (VMs).
Storage Protocol	Fibre Channel (FC), iSCSI, NFS, SMB — transport data between compute and storage.
Bottleneck	Storage I/O must traverse the network, adding latency. Scaling requires forklift upgrades.

Key Challenges of the Three-Tier Model

Complex management: 3 separate teams, 3 separate toolsets, 3 separate upgrade cycles
High CapEx: standalone SAN, network switches, and servers bought separately
Poor scalability: scale-up storage arrays hit physical/cost limits
Single points of failure: centralized storage is a critical dependency
Performance bottleneck: all I/O travels across shared SAN fabric
Long procurement cycles: months to deploy new capacity

1.2 The HCI Revolution

Hyper-Converged Infrastructure (HCI) collapses the three tiers into a single software-defined platform running on commodity x86 nodes. Compute, storage, and networking are all integrated into one appliance managed by a unified software stack.

HCI	Hyper-Converged Infrastructure — collapse compute, storage, and networking into a single x86 node cluster.
Software-Defined Storage (SDS)	Storage services (replication, dedup, compression) implemented in software, not dedicated hardware.
Distributed Storage Fabric	Nutanix DSF — the storage layer that pools all local drives across nodes into a single logical datastore.
Scale-Out	Add capacity and performance by adding nodes — no forklift upgrades required.
Controller VM (CVM)	Nutanix software VM running on each node, managing local storage and serving as storage endpoint for the hypervisor.

HCI Benefits vs. Three-Tier

Simplified operations: single pane of glass via Prism UI
Linear scalability: add nodes to expand capacity and performance simultaneously
Reduced CapEx/OpEx: commodity hardware, no separate SAN, fewer network components
Faster deployment: nodes image in minutes; clusters form in 30–45 min
Built-in resiliency: data distributed across nodes with configurable redundancy

1.3 Nutanix Company & Platform Overview

Founded in 2009, Nutanix pioneered enterprise HCI. The platform has evolved from pure HCI storage into a full cloud platform (cloud-native, DR, files, networking, databases, VDI).

AOS	Acropolis Operating System — the Nutanix storage and hypervisor management OS running on each node.
AHV	Acropolis Hypervisor — Nutanix's own KVM-based hypervisor, included at no extra cost.
DSF	Distributed Storage Fabric — pools node-local SSDs and HDDs into a distributed datastore.
Prism Element (PE)	Per-cluster management UI — cluster-level operations, VM management, monitoring.
Prism Central (PC)	Multi-cluster management plane — manage thousands of clusters from a single UI.
NCC	Nutanix Cluster Check — automated health check and diagnostics framework.
Foundation	Nutanix imaging tool — used to bootstrap bare-metal nodes and create clusters.
Pulse	Optional telemetry service that sends diagnostic data to Nutanix support.

📘

Key ConceptNutanix separates the data plane (AOS/DSF) from the management plane (Prism). Data continues to flow even if Prism is temporarily unavailable.

SECTION 02

Understanding Physical Construct & Basic Behavior

2. Physical Construct & Basic Behavior

2.1 Node Anatomy

A Nutanix node is a standard 1U or 2U x86 server containing CPU, RAM, storage devices (NVMe/SSD/HDD), and a NIC. Each node runs a hypervisor and the CVM (Controller VM).

Node	Single physical server — the fundamental building block of a Nutanix cluster.
Block	A chassis holding 1, 2, or 4 nodes. NX-1175S-G9 = 1-node; NX-3155G-G9 = 2-node; NX-3060-G9 = 4-node.
CPU	Intel Xeon Scalable or AMD EPYC CPUs — multiple sockets supported.
RAM	DDR4/DDR5 DIMMs — directly mapped to host/CVM/VMs.
NVMe SSD	Fastest tier — PCIe-attached SSDs for hot data and metadata. Also called 'Tier 0'.
SSD (SATA/SAS)	Performance tier — warm data. Used in all-flash or hybrid nodes.
HDD	Capacity tier — cold data in hybrid nodes (SSD+HDD config).
NIC	Network Interface Card — 10/25/40/100 GbE for VM traffic, storage replication, and management.
IPMI/BMC	Baseboard Management Controller — out-of-band management (remote KVM, power control).
Boot Drive	Dedicated SSD/HDD/M.2 or SD card for the hypervisor OS — separate from data drives.

Node Storage Tiers

Tier 0 — NVMe SSDs: Ultra-low latency metadata and hot data, CVM memory-tiered
Tier 1 — SATA/SAS SSD: Active working set, read cache (Extent Cache), warm data
Tier 2 — HDD (hybrid only): Bulk capacity for cold data, sequential workloads

2.2 Cluster Architecture

A Nutanix cluster is a collection of 3 or more nodes (minimum 3 for redundancy) sharing a distributed storage pool. All nodes communicate over standard Ethernet.

Minimum Cluster Size	3 nodes (required for Replication Factor 2 — RF2).
Cluster IP	Virtual IP (VIP) representing the cluster management endpoint — floats among CVMs.
Data Services IP	iSCSI target IP used by external initiators (for guest-initiated iSCSI connections).
Node Quorum	Majority of nodes must be online for cluster operations. Losing >50% = cluster fenced.
Zookeeper	Distributed coordination service running on CVMs — manages cluster state and configuration.
Medusa	Nutanix metadata database (key-value store) — maps logical blocks to physical locations.
Cassandra	Distributed NoSQL DB used by Medusa for metadata persistence across CVMs.

2.3 Controller VM (CVM) Deep Dive

The CVM is a virtual machine that runs on every node, providing all storage I/O services. It intercepts storage requests from the hypervisor and fulfills them locally or via the distributed fabric.

CVM Role	Manages local storage devices, serves iSCSI/NFS to hypervisor, handles replication, runs metadata services.
CVM Resources	4–32 vCPUs, 12–72 GB RAM (varies by AOS version and workload) — reserved exclusively for CVM.
CVM IP	Static IP on the storage network — all hypervisor storage traffic is directed here.
Passthrough	Physical NVMe/SSD/HDD devices are passed through directly to CVM (not shared with hypervisor).
CVM HA	If CVM on one node fails, hypervisor redirects storage I/O to CVMs on other nodes automatically.
CVM Tunnel	When local CVM is down, a backplane network path (192.168.5.x) routes I/O to other CVMs.

⚠️

WarningNever power off a CVM without first putting it into maintenance mode. Abruptly stopping the CVM disrupts local storage I/O and triggers remote I/O paths.

2.4 Key CVM Services

Stargate	Primary I/O manager — handles all read/write operations, caching, deduplication, compression, erasure coding.
Curator	Background housekeeping — runs MapReduce jobs for dedup, compression, tiering, disk balancing, and garbage collection.
Medusa	Metadata abstraction layer — fronts Cassandra, provides consistent key-value API for all metadata lookups.
Cassandra	Persistent metadata ring — stores vDisk extents, block mappings, and cluster topology data.
Zookeeper	Cluster state management — leader election, distributed lock, configuration store.
Cerebro	Replication and DR engine — manages protection domains, snapshots, async/sync replication.
Acropolis	AHV management service — handles VM lifecycle, network, and storage attachment for AHV VMs.
Prism Gateway	Web application server for Prism Element UI — proxies API calls to underlying services.
Genesis	Cluster bootstrap service — initializes cluster on first start, manages CVM startup/stop.
Mantle	Security service — key management, encryption at rest operations.

SECTION 03

Logical Construct & Detailed Behavior

3. Logical Construct & Detailed Behavior

3.1 Distributed Storage Fabric (DSF)

DSF presents all node-local storage as a single unified storage pool. VMs see a logical datastore (NFS mount on AHV, VMware, or Hyper-V), while DSF handles data placement, protection, and tiering transparently.

Container	Logical storage pool presented to the hypervisor as a datastore. Maps to a subset (or all) of the Storage Pool.
Storage Pool	Aggregate of all physical storage across all nodes in the cluster.
vDisk	Virtual disk file (.vhdx / .vmdk / .qcow2) stored in a container. The disk image a VM uses.
Extent	Fixed 1 MB chunk — basic unit of I/O in Nutanix. Each write is broken into extents.
Extent Group	4 MB collection of extents — the unit of replication and deduplication.
Oplog	Write buffer — volatile writes go here first (SSD-backed), then destaged to persistent storage.
Extent Store	Persistent tier on SSD/HDD — final resting place for all data after Oplog destage.
Unified Cache	Flash-based read cache in Stargate — caches hot extents for low-latency reads.

3.2 I/O Path — Write Flow

VM issues a write → NFS/iSCSI write request to local CVM

CVM Stargate receives the write and acknowledges the guest VM

Data is written to the local Oplog (SSD-backed) for low-latency response

Simultaneously, data is replicated to remote CVMs' Oplogs (RF2=1 copy, RF3=2 copies)

Write is acknowledged to the VM after replication is confirmed

Curator destages Oplog data to Extent Store (SSD/HDD) during background operations

Metadata in Cassandra is updated to reflect final data location

📘

Key ConceptNutanix uses a 'write locally, replicate immediately' model. The local Oplog provides fast write acknowledgment while async destage moves data to persistent storage.

3.3 I/O Path — Read Flow

VM issues a read → NFS/iSCSI read request to local CVM

Stargate checks Unified Cache — if hit, data returned immediately (~100 μs)

If cache miss, Stargate checks local SSD Extent Store

If not local, request forwarded to the CVM holding the data replica

Data returned to requesting VM

Hot data gets promoted to Unified Cache for future reads

3.4 Data Locality & I/O Preference

Data Locality	Data belonging to a VM is preferentially stored on the node running that VM.
Remote I/O	When data is not local (e.g., after VM migration), Stargate fetches from the remote node holding the replica.
Intelligent Tiering	Curator automatically moves hot data to SSD and cold data to HDD based on access patterns.
ILM	Information Lifecycle Management — the background process governing data movement between tiers.
Oplog Draining	When Oplog reaches threshold, Curator drains it to Extent Store. Large sequential writes bypass Oplog entirely.

3.5 Container & vDisk Configuration

Compression	Container-level setting — inline or post-process. Reduces storage consumption.
Deduplication	Container-level setting — removes duplicate data blocks. Effective for VDI/clone workloads.
Erasure Coding	Container-level setting — replaces full copies with parity chunks. Reduces storage at the cost of CPU.
RF2	Replication Factor 2 — 1 primary copy + 1 replica. Survives 1 node failure. Default.
RF3	Replication Factor 3 — 1 primary + 2 replicas. Survives 2 node failures. Used for mission-critical VMs.
Reserved Capacity	Minimum guaranteed capacity for a container — prevents starvation in multi-container clusters.
Advertised Capacity	Maximum capacity a container can present to the hypervisor (includes logical savings).

SECTION 04

How Data Resiliency Is Managed by Nutanix

4. Data Resiliency

4.1 Replication Factor (RF)

RF2 (Default)	Data stored on 2 nodes. Tolerates failure of 1 node or 1 disk. Minimum cluster size: 3 nodes.
RF3	Data stored on 3 nodes. Tolerates failure of 2 nodes simultaneously. Minimum cluster size: 5 nodes.
Metadata RF	Metadata (Cassandra) uses RF3 regardless of data RF — ensures metadata consistency.
Replica Placement	Replicas never placed on the same node, same block (chassis), or same rack (if rack-aware).
Block Awareness	Nutanix places replicas across different chassis (blocks) to survive block-level failures.
Rack Awareness	For deployments with rack diversity — replicas placed across different racks.

4.2 Resilience Domains & Failure Tolerance

Node Failure	CVM stops. Hypervisor routes storage I/O to remote CVM replicas. VMs continue running (brief I/O pause).
Disk Failure	Failed disk detected by Stargate. Affected extents rebuilt from replicas on surviving nodes.
Block Failure	Entire chassis fails. Block-aware RF2 still provides access via replicas on other blocks.
Network Partition	Zookeeper quorum determines cluster health. Minority partition fences itself to prevent split-brain.
CVM Failure (node alive)	Hypervisor redirects I/O via backplane tunnel to another CVM. VMs continue without interruption.
Metro Failure	With Sync Rep/Metro Availability configured, secondary site takes over automatically.

📘

Key ConceptIn RF2, losing 1 node puts the cluster in a 'degraded but available' state. Data is still accessible from the remaining replica, and Curator immediately begins rebuilding lost replicas.

4.3 Rebuild & Self-Healing

Failure detected (hardware, heartbeat timeout, Zookeeper quorum)

Medusa marks affected extents as 'under-replicated'

Curator schedules MapReduce jobs to rebuild missing extents

Surviving replicas are used to reconstruct data on remaining healthy nodes

Data is redistributed to balance capacity across surviving nodes

Cluster returns to full redundancy — no manual intervention required

4.4 Disk Balancing

Disk Balancing	Background process that migrates data extents to equalize usage across all disks and nodes.
Trigger	Triggered when any node's disk usage deviates >15% from cluster average, or on topology change.
Impact	Low-priority background I/O — does not impact foreground VM workloads significantly.

4.5 Erasure Coding (EC-X)

Erasure Coding	Breaks data into k data chunks and m parity chunks. Can reconstruct from any k chunks of k+m.
EC-X Modes	3+1 (3 data, 1 parity — 1.33x overhead), 4+1 (1.25x overhead), 4+2 (1.5x overhead, survives 2 failures)
When Applied	EC-X is applied to cold data by Curator during background operations — not to hot/active data.
Hot Data	Hot/active data remains in RF2/RF3 full copies for lowest read latency; EC only on cold extents.
Decode Overhead	Read of EC data requires reconstruction if a stripe is unavailable — slight read latency increase.
Space Savings	EC 4+1 reduces overhead from 100% (RF2) to 25% — significant capacity savings at scale.

SECTION 05

Space Optimization Features & Impact on Data Path

5. Space Optimization Features

5.1 Compression

Inline Compression	Data is compressed before being written to disk. Saves space immediately but adds CPU overhead on write path.
Post-Process Compression	Data written uncompressed, then Curator compresses during background operations. Lower write latency.
Compression Algorithm	LZ4 (fast, lower ratio) or Zlib/Deflate (higher ratio, more CPU). LZ4 default for performance.
Typical Savings	30–60% reduction for most VM workloads. Already-compressed data (media, zip) sees minimal benefit.
Container Level	Configured per container — can enable for some workloads, disable for others (e.g., databases).

5.2 Deduplication

Inline Dedup	Duplicate detection before write. Fingerprinting performed in-memory — requires significant RAM.
Post-Process Dedup	Curator identifies duplicates after data lands on SSD. More common and less RAM-intensive.
Fingerprinting	SHA-1 hash computed for each extent. Matching fingerprints indicate duplicate data blocks.
Best Use Case	VDI full clones, linked clones, similar OS images — where blocks repeat across many VMs.
Dedup Ratio	Varies widely — VDI: 4–10x, mixed workloads: 1.2–2x, unique data (databases): ~1x.
RAM Requirement	Dedup requires additional CVM RAM for fingerprint tables — plan accordingly.

5.3 Thin Provisioning

Thin Provisioning	vDisks are created with a logical size larger than physical allocation. Space allocated as data is written.
Thick Provisioning	Full logical capacity pre-allocated — no overcommit. Used for latency-sensitive workloads.
Overcommit	Present more logical storage than physical. Relies on compression/dedup savings to satisfy workloads.
Space Reclamation	When guest deletes files, Nutanix reclaims the physical space via TRIM/UNMAP operations.

5.4 Feature Impact Comparison

Feature	Write Impact	Read Impact	CPU Impact	Best Workload
Inline Compression	Higher latency	Lower latency	Medium	Compressible data
Post-Process Compression	Minimal	Minimal	Low (background)	General purpose
Inline Dedup	Higher latency	Lower latency	High	VDI clones
Post-Process Dedup	Minimal	Minimal	Medium (background)	Large clone farms
Erasure Coding	No impact (cold only)	Slight increase	Low–Medium	Cold/archive data
Thin Provisioning	Minimal	Minimal	Minimal	General purpose

⚠️

WarningDo NOT enable inline dedup on databases or workloads with unique data. The fingerprinting overhead will reduce write performance with no space benefit.

SECTION 06

6. Networking Concepts

6.1 Nutanix Network Architecture

Nutanix uses standard Ethernet (10/25/40/100 GbE) for all traffic — no specialized SAN fabric required. All network functions are software-defined through the hypervisor.

IPMI Network	Out-of-band management for BMC — used for remote KVM, power control, Foundation imaging.
CVM Network	Storage network — CVM IPs communicate with each other for replication and metadata sync.
Host Network	Hypervisor management — AHV host IP, VMware ESXi management IP.
VM Network	Guest VM traffic — VLAN-tagged or untagged network attached to VM vNICs.
Backplane (192.168.5.x)	Internal CVM-to-host tunnel — used when CVM redirects I/O to remote CVM.
Cluster VIP	Virtual IP floating among CVMs — represents the cluster management endpoint.
Data Services IP	iSCSI target IP — for external initiators connecting block volumes.

6.2 AHV Networking

Open vSwitch (OVS)	Software virtual switch used in AHV — provides L2 switching, VLAN tagging, bonding.
Bridge (br0, br1…)	OVS bridge — logical L2 domain connecting host interfaces and VM vNICs.
Uplink	Physical NIC port(s) connected to the OVS bridge for upstream connectivity.
Bond	Combines multiple physical NICs for redundancy and/or bandwidth aggregation.
Active-Backup Bond	One NIC active, others standby. Failover only — no bandwidth aggregation.
Active-Active (LACP)	All NICs carry traffic simultaneously. Requires LACP on upstream switch.
VLAN	802.1Q VLAN tagging on trunk uplinks — allows multiple L2 networks on one physical uplink.
vNIC	Virtual NIC attached to a VM — connected to a specific OVS bridge/VLAN.
Network (AHV)	Logical network object in Prism — defines VLAN ID, IP pool (optional), and DHCP settings.

6.3 VLAN Configuration

VLAN 0 (untagged): Native VLAN — packets have no 802.1Q tag
Tagged VLAN: Specify VLAN ID 1–4094 when creating a network in Prism
Trunk port: Switch port passes multiple VLANs to the node — required for tagged VMs
Access port: Switch port passes only one VLAN — used for management or single-VLAN hosts

6.4 MTU & Jumbo Frames

Standard MTU	1500 bytes — default Ethernet frame size.
Jumbo Frames	MTU 9000 bytes — reduces CPU overhead for large I/O transfers. Recommended for storage traffic if switch supports it.
CVM MTU	Should match the physical switch configuration — mismatched MTU causes silent packet drops.
Flow Control	802.3x pause frames — can cause congestion. Nutanix recommends Priority Flow Control (PFC) for NVMe-oF.

SECTION 07

Understanding Concepts of Native Disaster Recovery

7. Native Disaster Recovery

7.1 DR Concepts & Terminology

RPO	Recovery Point Objective — maximum acceptable data loss (e.g., RPO=1hr means lose up to 1hr of data).
RTO	Recovery Time Objective — maximum time to restore service after a disaster.
Protection Domain (PD)	Logical grouping of VMs and vDisks for DR purposes — configured replication schedules.
Snapshot	Point-in-time copy of VM data — can be local (for restore) or remote (for DR).
Async Replication	Snapshots taken at scheduled intervals and replicated to remote site. RPO = snapshot interval.
Sync Replication	Every write replicated synchronously to remote site before acknowledgment. RPO=0.
Near-Sync	Replication every 1–15 minutes using lightweight change tracking. Sub-minute RPO options.
Failover	Activating VMs at the DR site — planned (graceful) or unplanned (disaster recovery).
Failback	Returning VMs to the original site after DR event is resolved.
Test Failover	Non-disruptive test of DR capabilities without impacting production.

7.2 Protection Policies

Consistency Group	Set of VMs that must be snapshotted together (crash-consistent) — ensures app-consistent recovery.
Schedule	Defines snapshot frequency, retention count, and remote site replication target.
Remote Site	A paired Nutanix cluster where replicas are sent. Configured in Prism Data Protection.
Retention	How many snapshots are kept — local and remote retention can differ.
Network Mapping	Maps source VM networks to target networks at DR site — applied during failover.
VM Override	Per-VM custom settings that override Protection Domain defaults.

7.3 Async vs. Sync Replication Comparison

Aspect	Async Replication	Sync Replication (Metro)
RPO	Minutes to Hours	Zero (RPO=0)
RTO	Minutes to Hours	Sub-minute
Distance	Unlimited (WAN)	<5ms RTT (~300 miles)
Write Impact	Minimal (async)	~2x write latency
Use Case	DR, backup, compliance	Always-on HA, Metro clusters
Nutanix Product	AOS DR (Cerebro)	Sync Rep / Leap

7.4 Nutanix Leap (Cloud DR)

Leap	Nutanix DR-as-a-Service — orchestrates failover/failback between on-premises clusters and cloud (AWS, Azure).
Recovery Plan	Orchestrated runbook defining VM boot order, network mappings, and IP changes during failover.
Failover Steps	1. Trigger failover 2. Apply network mappings 3. Boot VMs per recovery plan 4. Validate
Test Failover	Brings up VMs at DR site on an isolated network — non-disruptive, validates RPO/RTO.
Prism Central Required	Leap is managed from Prism Central — requires both source and target registered in PC.

SECTION 08

Nutanix Files, Self-Service, Flow & Leap

8. Nutanix Software Products

8.1 Nutanix Files

Nutanix Files is a scale-out software-defined NAS solution deployed as a cluster of File Server VMs (FSVMs) on top of Nutanix.

FSVM	File Server VM — Nutanix-managed VM providing NFS and SMB file services. 3+ FSVMs per Files cluster.
SMB	Server Message Block — Windows file sharing protocol. Nutanix Files supports SMB 2.0/3.0.
NFS	Network File System — Linux/Unix file sharing. Nutanix Files supports NFS v3 and v4.
Share	Logical file share exposed to clients — can be SMB or NFS. Distributed across FSVMs.
Distributed Share	Files auto-balances data and load across all FSVMs — no single FSVM bottleneck.
Home Share	Each user gets a dedicated home directory. Used for end-user file storage.
General Share	Shared folder accessible by multiple users — departmental shares, application data.
File Analytics	Optional add-on — provides audit trails, anomaly detection, ransomware protection for Files data.
Scale-Out	Add FSVMs to expand capacity and throughput without downtime.
Integration	Files storage is backed by Nutanix DSF — inherits all storage efficiency features.

8.2 Nutanix Self-Service (Calm)

Nutanix Self-Service (SSP/Calm) is an application automation and orchestration platform that enables DevOps-style self-service provisioning of complex multi-cloud applications.

Blueprint	Template that defines an application's infrastructure, services, configuration, and lifecycle actions.
Application	Deployed instance of a Blueprint — running with its full service topology.
Service	Individual VM or group of VMs within an Application (e.g., web tier, DB tier).
Action	Lifecycle operation defined in a Blueprint (Create, Delete, Start, Stop, Scale Out, Scale In, custom).
Task	Individual step within an Action — can be shell script, PowerShell, HTTP, Nutanix, or VMware task.
Variable	Parameterized input defined in Blueprint — can be static, dynamic, or user-entered at launch.
Provider	Cloud target where Services are provisioned — Nutanix, AWS, Azure, GCP, VMware, K8s.
Marketplace	Self-service catalog — published Blueprints available for users to launch with one click.
Project	Tenant isolation unit in Calm — defines users, roles, resource limits, and providers.
Policy Engine	Approval workflows, quota enforcement, and governance rules for Self-Service requests.
NuCalm (Runbook)	Workflow automation for Day-2 operations without VM deployment — IT process automation.

8.3 Nutanix Flow (Microsegmentation)

Nutanix Flow provides distributed, application-centric firewall policies implemented directly in the AHV OVS layer. No dedicated hardware appliance required.

Microsegmentation	Zero-trust security model — each VM's traffic is controlled by its own security policy at the vNIC level.
Security Policy	Set of rules defining allowed/denied traffic for a group of VMs. Applied by OVS on each host.
Application Security Policy	Policy scoped to an application — defines inbound/outbound rules per tier (web, app, db).
Quarantine Policy	Isolates a potentially compromised VM — blocks all traffic except forensics/ICMP.
Isolation Policy	Prevents two groups of VMs from communicating — e.g., dev and prod isolation.
Category	Label applied to a VM (e.g., AppType=Web, Env=Prod). Policies target categories, not individual VMs.
Service Chain	Redirect traffic through a third-party virtual network appliance (firewall, IDS) via policy.
Allow List Mode	Deny all by default — only explicitly permitted traffic flows. Most secure posture.
Monitor Mode	Policy evaluated but not enforced — used to audit traffic before enforcing.
Flow Visualization	Prism Central Traffic Visualization — shows network flows between VMs as an interactive graph.

8.4 Nutanix Leap — Recovery Plan Details

Recovery Plan	Defines VM boot order (stages), network mappings, IP address changes, and scripts.
Stage	Group of VMs that boot together. Multiple stages enforce boot ordering (infra → app → web).
IP Mapping	Maps production VM IPs to DR site IPs — can be 1:1, subnet change, or dynamic.
Script	Custom PowerShell or shell scripts run during failover — DNS update, load balancer reconfiguration.
Status Checks	Leap validates VM state, replication health, and network readiness before allowing failover.

SECTION 09

Introduction to Foundation & Node Imaging

9. Foundation & Node Imaging

9.1 What is Foundation?

Foundation is the Nutanix tool used to image (install) bare-metal nodes with the chosen hypervisor (AHV, ESXi, Hyper-V) and AOS version, and then create or expand clusters.

Foundation VM	A pre-built VM (or physical laptop image) that hosts the Foundation application and web UI.
Foundation App	Web-based wizard that discovers nodes via IPMI and orchestrates the imaging process.
IPMI Discovery	Foundation connects to nodes via their IPMI/BMC IP to perform remote operations.
Hypervisor Bundle	ISO or tarball containing the hypervisor (AHV.iso, VMware ESXi ISO) to install.
AOS Bundle	Nutanix AOS software package (tar.gz) containing CVM image and AOS components.
Cluster Creation	After imaging, Foundation creates the cluster — sets node IPs, cluster name, cluster VIP.
Phoenix	Minimal Linux OS that boots on nodes during imaging — used to receive and install Foundation payloads.

9.2 Foundation Process — Step by Step

Pre-requisites: Network (IPMI, host, CVM IPs planned), DNS, NTP, hardware racked and powered

Deploy Foundation VM on a laptop or management server on the IPMI network

Open Foundation web UI (http://<foundation-vm-ip>:8000)

Discover nodes via IPMI — provide IPMI IP range and credentials

Select AOS version and hypervisor type/version

Upload or provide URL to AOS and hypervisor bundles

Assign node IPs (IPMI, Hypervisor Host, CVM) and cluster parameters (name, VIP, NTP, DNS)

Run Foundation — nodes boot Phoenix, receive image payload, install, configure, reboot

Cluster forms — CVMs negotiate, Zookeeper quorum establishes, Prism becomes available

Access Prism Element at https://<cluster-VIP>:9440

9.3 Foundation Modes

Normal Foundation	Full imaging — wipes existing OS and installs fresh hypervisor and AOS.
Re-Foundation	Re-image an existing Nutanix node — data will be DESTROYED. Used for recovery.
Add Node	Image a new node and add it to an existing cluster without disrupting running VMs.
1-Click Foundation	Automated discovery and imaging with minimal input — uses DHCP for bootstrap.
Field Foundation	Foundation from a USB key — for environments with no network-based Foundation VM.
Foundation Central	Centralized Foundation management for imaging nodes at remote locations at scale.

9.4 Foundation Requirements

IPMI (BMC) access: Foundation VM must reach all node IPMI IPs
L2 connectivity: Foundation VM, IPMI, and host management network on same broadcast domain or routable
Correct AOS + hypervisor bundle: versions must be compatible (AOS Compatibility Matrix)
DNS and NTP: required for cluster formation and ongoing operations
Time sync: All nodes and Foundation VM must have synchronized time (<5 min difference)

⚠️

WarningFoundation destroys all existing data on the target nodes. Always verify the correct IPMI IPs are selected before starting the imaging process.

SECTION 10

Portal, Test Drive, University & Certification

10. Nutanix Resources

10.1 Nutanix Portal (support.nutanix.com)

Software Downloads	AOS, AHV, Foundation, NCC, Prism Central, Files, Flow, Calm packages.
Knowledge Base (KB)	Searchable articles for known issues, upgrade procedures, and troubleshooting guides.
Support Cases	Open, track, and update support cases. Attach logs (log bundles) to expedite resolution.
NCC Checks	NCC plugin reference — understand what each health check validates.
Compatibility Matrix	AOS ↔ AHV ↔ Prism Central ↔ third-party software version compatibility.
Field Alerts	Proactive notifications for critical issues, required upgrades, and hardware advisories.
License Management	View, allocate, and manage Nutanix software licenses (AOS, Prism Pro, Files, etc.).
Log Collection	Guides for collecting logs via Prism or ncli for support case attachment.

10.2 Nutanix Test Drive

Hosted Labs	Pre-provisioned Nutanix clusters in Nutanix's data center — access via browser.
Available Products	AHV, Prism Element, Prism Central, Files, Flow, Calm/Self-Service — product-specific labs.
Duration	60–90 minute sessions with guided walkthroughs.
Use Case	Pre-sales evaluation, learning, demo preparation — no commitment required.
No Hardware Required	Access via web browser — use RDP or direct browser-based console.

10.3 Nutanix University & Certification

NCP-MCI	Nutanix Certified Professional — Multicloud Infrastructure. Entry-level administrator cert.
NCP-DB	NCP — Database Automation. Focuses on Era (database management).
NCP-VDI	NCP — Virtual Desktop Infrastructure. Frame and AHV VDI deployments.
NCX	Nutanix Certified Expert — advanced architecture and design. Requires NCP prerequisite.
NPX	Nutanix Platform Expert — highest cert level. Requires NCX + design challenge.
Free Courses	Nutanix University offers many free online courses — 'The Nutanix Bible', product courses.
Exam Vouchers	Purchased through Nutanix University — proctored via Pearson VUE or online proctoring.
Learning Path	Recommended: Free Foundation → NCP study → Practice labs → Exam.

SECTION 11

Prism Element UI Dashboard & Operations

11. Prism Element (PE) UI

11.1 Prism Architecture

Prism Element (PE)	Per-cluster management UI — runs on the cluster itself. URL: https://<cluster-VIP>:9440
Prism Central (PC)	Multi-cluster management — deployed separately. URL: https://<PC-IP>:9440
Prism Pro	License tier adding advanced analytics, automation, and X-Play capabilities to PC.
REST API	All Prism operations available via REST API v2 (PE) and v3 (PC). API Explorer at /api/nutanix/v2.0
Session Timeout	Default 15 minutes of inactivity. Configurable in Settings.
Default Credentials	Username: admin \| Password: set during cluster creation (Foundation).
SSL Certificate	Self-signed by default. Replace with CA-signed cert via Settings > SSL Certificate.

11.2 Main Dashboard Widgets

Cluster Health Widget	NCC health score — green/yellow/red. Click to drill into health checks.
Hardware Widget	Block diagram showing node/disk state — color coded by health status.
Storage Summary	Total capacity, used, available, savings from dedup/compression.
Performance Widget	Real-time IOPS, throughput (MB/s), and latency charts for the cluster.
VM Summary	Count of VMs by state (running/stopped). Click to go to VM list.
Alerts Widget	Recent critical/warning alerts with quick links to investigation.
Data Resiliency Status	Shows current RF and if any data is under-replicated.

11.3 Navigation Menu Reference

Home	Main dashboard — overview widgets as described above.
VM	Virtual machine list — create, power on/off, clone, migrate, delete VMs.
Storage	Container, vDisk, and Storage Pool management.
Network	AHV network configuration — create/edit networks (VLAN), IP address management.
Data Protection	Protection Domains, replication schedules, remote sites, snapshots.
Hardware	Physical node/disk view — disk health, node status, alerts.
Analysis	Performance analysis charts — custom time range, custom metrics.
Alerts	All cluster alerts — filter by severity, category, acknowledge, resolve.
Tasks	Background task tracker — imaging, snapshot, replication, upgrade tasks.
Health	NCC health check results — individual check results and recommendations.
Settings	Cluster configuration — NTP, DNS, SMTP, SSL, authentication, licensing.

SECTION 12

Hands-On CUD Operations (Create, Update, Delete)

12. CUD Operations

12.1 VM Lifecycle — Create

Navigate to VM > Create VM

Enter VM Name, description

Set vCPU count, number of cores per vCPU, and memory (GiB)

Add disk: Click '+ Add New Disk' → Clone from Image Service or Empty disk → set size and bus type (SCSI/IDE)

Add NIC: Click '+ Add New NIC' → select network → VLAN will be assigned

Optional: Enable Guest Customization (cloud-init for Linux / Sysprep for Windows)

Click Save → VM appears in list in 'Off' state

Power on: Select VM → Actions > Power On (or click power button)

12.2 VM Lifecycle — Update

Hot-add vCPU/RAM (AHV): Some guests support hot-add without reboot — check guest OS support
Add disk: VM > select VM > Update > + Add New Disk → save (live on AHV)
Resize disk: Update vDisk size from VM Update screen (expand only — no shrink)
Change network: Remove and re-add NIC or edit existing NIC VLAN assignment
Rename VM: Update VM name — does not affect guest hostname
Update CPU/RAM offline: Power off VM first if hot-add not supported

12.3 VM Lifecycle — Delete

Ensure VM is powered off

Select VM → Actions > Delete

Optionally check 'Delete all snapshots' to remove associated protection snapshots

Confirm deletion — this is irreversible if snapshots are not retained

⚠️

WarningDeleting a VM only deletes the VM object and its vDisks. Independent vDisks or Images in the Image Service must be deleted separately.

12.4 Image Management

Image Service	Nutanix central repository for disk images (ISO, qcow2, vmdk) — used as templates for VM creation.
Upload Image	Settings > Image Configuration > + Upload Image → provide URL or file upload.
Image Types	DISK (bootable vDisk template), ISO (CD-ROM for OS installation).
Clone from Image	When creating a VM disk, select 'Clone from Image Service' → fastest way to deploy VMs.
Delete Image	Settings > Image Configuration → select image > Delete. Frees storage used by the image.

12.5 Container Management

Create Container: Storage > + Storage Container → name, select storage pool, set RF, enable/disable features
Edit Container: Select container > Update → modify compression, dedup, EC settings
Delete Container: Only possible if empty (no vDisks). Migrate VMs first.
Mount Point: Container is auto-mounted on all AHV hosts as an NFS export at /var/lib/containers/<name>

SECTION 13

Day 2 Operations

13. Day 2 Operations

13.1 Cluster Expansion (Add Node)

Rack and cable new node(s), assign IPMI IP

Use Foundation to image the new node(s) with the same AOS/hypervisor version

In Foundation, select 'Expand Cluster' and target the existing cluster

Foundation images the node and adds it to the cluster automatically

Curator runs disk balancing to redistribute data to the new node

13.2 One-Click Upgrades (LCM)

LCM	Life Cycle Manager — automated software inventory and upgrade framework in Prism.
Inventory	LCM scans the cluster and identifies current versions and available upgrades.
Upgrade Order	LCM follows safe upgrade order: NCC → Foundation → AOS → AHV → Firmware.
Rolling Upgrade	One node at a time — VMs live-migrated off node, node upgraded, node rejoins cluster.
Dark Site	LCM can operate without internet — upload upgrade bundles to LCM dark site server.
Pre-check	NCC pre-upgrade health check runs before every upgrade — blocks upgrade if critical issues found.
AHV Upgrade	Done via Prism Settings > AHV Upgrade OR LCM. Requires at least 2 nodes for live migration.

13.3 VM Live Migration

Manual: VM > Select VM > Migrate > choose target host
Automatic: ADS (Acropolis Dynamic Scheduler) automatically balances VM placement
During node maintenance: put node in maintenance mode — all VMs auto-migrate off

13.4 Host Maintenance Mode

Prism: Hardware > select node > Enter Maintenance Mode

All running VMs live-migrate to other nodes

CVM on the node enters standby — I/O path redirected to cluster

Perform maintenance (disk replacement, hardware repair, upgrade)

Exit Maintenance Mode — CVM restarts, node rejoins cluster, VMs can migrate back

13.5 Capacity Planning & Analysis

Runway Report	Prism Pro feature — projects when cluster will run out of storage or compute based on trends.
What-If Analysis	Model adding nodes or VMs and see projected impact on cluster resources.
Storage Efficiency	Monitor overall savings from compression/dedup/EC in the Storage Summary widget.
Anomaly Detection	Prism Pro — detects unusual VM behavior and alerts on performance deviations.

SECTION 14

Introduction & Configuration of Data Protection

14. Data Protection Configuration

14.1 Protection Domain Setup

Navigate to Data Protection > Protection Domains

Click + Create Protection Domain > Async DR (or Metro for sync)

Name the Protection Domain

Add VMs or vDisks to the Protection Domain (Entities tab)

Create Consistency Groups — group VMs that must snapshot together

Set Schedule: Click + Add a Schedule → define RPO, local retention, remote site, remote retention

Optionally define Remote Site (if not already configured)

Save — Protection Domain is active and will take snapshots per schedule

14.2 Remote Site Configuration

Data Protection > Remote Sites > + Remote Site

Enter remote cluster name, remote CVM/cluster IPs

Enter remote cluster credentials

Define bandwidth throttling if needed (limit replication bandwidth during business hours)

Optionally configure network mapping (source VLAN → target VLAN)

Save and validate connectivity

14.3 Snapshot Types

Crash-Consistent	Point-in-time copy — VM is not quiesced. Similar to pulling power. Safe for most workloads.
Application-Consistent	Guest OS is notified via VSS (Windows) or pre/post scripts. Flushes I/O buffers for clean state.
Local Snapshot	Stored on the local cluster — for fast restore. Doesn't protect against site failure.
Remote Snapshot	Replicated to remote site — protects against site-level failure.
Ad-hoc Snapshot	Manual snapshot outside of scheduled policy — taken on demand.

14.4 Restore from Snapshot

Data Protection > Table > select Protection Domain

Select the snapshot to restore from (by date/time)

Click Restore → choose: Restore to original location OR Clone to new VMs

For clone: provide new VM names and target container

Restored/cloned VMs appear in the VM list — power on to verify

SECTION 15

Health, Analysis, Alerts & Tasks in Prism Element

15. Health, Analysis, Alerts & Tasks

15.1 Health (NCC)

NCC	Nutanix Cluster Check — collects diagnostic data and runs checks against known best practices/issues.
Check Categories	Hardware, Software, Network, Cluster Services, Storage, Security, Configuration.
Severity Levels	PASS (green), INFO, WARNING (yellow), FAIL (red), ERROR.
Auto-Run	NCC runs automatically every 4 hours and on-demand via Prism or command line.
Run from Prism	Health > Run NCC > select checks > Run.
Run from CLI	ncc health_checks run_all (on any CVM via SSH).
Log Location	/home/nutanix/data/logs/ncc-output.log
Plugins	NCC checks are plugin-based — updated independently of AOS via Portal download.

15.2 Analysis (Performance Charts)

Add Chart	Analysis > + Add Chart → select entity type and metric.
Metrics	IOPS (read/write/total), Throughput (MB/s), Latency (ms), CPU %, Memory %, I/O queue depth.
Time Range	Last 1 hour to 30 days. Custom time range supported.
Entity Types	Cluster, Node, VM, Disk, Container, vDisk, Network.
Baseline	Prism Pro feature — shows expected performance range; highlights anomalies.
Export	Charts can be exported as PNG or data as CSV for reporting.

15.3 Alerts

Alert Types	Critical (requires immediate action), Warning (investigate soon), Info (informational).
Sources	NCC checks, hardware sensors (IPMI), software services, replication status, cluster state.
Acknowledge	Mark alert as seen — alert remains in list but is no longer shown as 'new'.
Resolve	Mark alert as resolved — removes from active alerts list.
Alert Policies	Settings > Alert Policies — customize thresholds, disable specific alerts.
Email Alerts	Settings > SMTP → configure mail server. Alert Policies → enable email notification.
SNMP Traps	Settings > SNMP → configure SNMP trap receiver for integration with monitoring systems.
Webhook	Prism Central → Alerts > Notification Policies → webhook for custom integrations.

15.4 Tasks

Task Monitor	Prism > Tasks — shows all running, completed, and failed background tasks.
Task Types	VM create/delete, snapshot, replication, upgrade, disk add, clone, image upload.
Task Status	Running (spinner), Succeeded (green), Failed (red), Queued (grey).
Task Details	Click any task to see step-by-step progress and error messages on failure.
API Task UUID	Each task has a UUID — can poll via REST API (/tasks/{uuid}) for automation.
Retention	Completed tasks visible in Prism for 30 days by default.

SECTION 16

Showcasing Failures & Impact on Cluster

16. Failure Scenarios & Impact Analysis

16.1 Single Disk Failure

Detection	Stargate/SMART detects disk failure → alert in Prism, disk marked offline.
Impact	Affected extents become under-replicated. Cluster remains fully operational.
Auto-Heal	Curator rebuilds missing copies from surviving replicas on other nodes/disks.
Time to Heal	Depends on amount of data on failed disk and cluster load — typically 1–4 hours.
Admin Action	Replace failed disk. Nutanix auto-detects new disk and uses it for rebalancing.

16.2 Single Node Failure

Detection	Node heartbeat timeout → Zookeeper removes node from quorum.
VM Impact	VMs on failed node: AHV restarts them on surviving nodes (HA). Brief downtime (~30 sec).
Storage Impact	Node's data replicas become unavailable → Curator rebuilds from surviving copies.
Data State	RF2: cluster in 'degraded' state — no further node failure tolerated until rebuilt.
Admin Action	Investigate hardware, replace node, use Foundation 'Expand Cluster' to re-add.

16.3 CVM Failure (Node Alive)

Detection	Hypervisor detects CVM heartbeat loss.
Impact	Local storage I/O redirected via backplane tunnel to other CVMs. Slight latency increase.
VM Impact	VMs on the affected node continue running — no HA event, no VM restart.
Auto-Recovery	CVM auto-restarts via hypervisor watchdog within 2–5 minutes.
Admin Action	Check CVM logs: /home/nutanix/data/logs/. Investigate service failure.

16.4 Network Failure

Switch Failure	With bonded NICs: traffic fails over to surviving NIC. VM network unaffected.
CVM Network Partition	Isolated CVM cannot replicate → Zookeeper removes from quorum. I/O rerouted.
Cluster Split-Brain	Minority partition fences itself — protects data integrity at cost of availability.
Management Network Loss	Prism UI unavailable but cluster and VMs continue I/O normally.

📘

Key ConceptPrism Element is a management plane — its unavailability does NOT affect the data plane. VMs continue to run and I/O continues even if Prism is down.

16.5 Cluster Shutdown & Startup Procedure

Graceful Shutdown

Notify users and coordinate maintenance window

Power off or live-migrate all guest VMs

SSH to any CVM → execute: cluster stop

Power off all nodes via IPMI or physical power

Cluster Startup

Power on all nodes

CVMs boot automatically with the hypervisor

Zookeeper quorum establishes — cluster comes online (3–5 minutes)

Verify cluster health: cluster status from CVM

Power on guest VMs

SECTION 17

Explaining & Configuring Options from Settings

17. Settings Configuration

17.1 Network Configuration

NTP Servers	Settings > NTP Servers — add NTP server IPs or FQDNs. Recommend 3+ servers for resilience.
DNS Servers	Settings > Name Servers — add DNS server IPs for hostname resolution.
Cluster Name	Settings > Cluster Details — cluster display name (does not change IPs).
Cluster VIP	Settings > Cluster Details — virtual IP for management. Can be changed (brief downtime).
SMTP	Settings > SMTP Server — configure mail relay for alert emails.
HTTP Proxy	Settings > HTTP Proxy — for clusters behind proxy needing internet access (LCM, Pulse).

17.2 Authentication & Security

Local Users	Settings > Local User Management — create, edit, delete local Prism user accounts.
Roles	Admin (full access), User (read-only). Custom roles available in PC.
Active Directory	Settings > Authentication > + New Directory — configure AD/LDAP for SSO.
SAML/SSO	Settings > Authentication > SAML — configure SAML 2.0 identity providers.
SSL Certificate	Settings > SSL Certificate — replace self-signed cert with CA-signed certificate.
Cluster Lockdown	Settings > Cluster Lockdown — disable SSH password auth, require key-based auth.
2FA (PC)	Prism Central supports two-factor authentication via TOTP (Google Authenticator).
Data-at-Rest Encryption	Settings > Data Encryption — enable Software Encryption or SEDs (hardware) with KMS.

17.3 Licensing

License Tiers	Starter (basic), Pro (analytics, policies), Ultimate (all features). Applied per cluster.
Apply License	Settings > Licensing — enter license key or connect to license server.
License Check	Settings > Licensing shows current license state and feature entitlements.
CE (Community Edition)	Free license for non-production/home lab use. Single-node or 3-node cluster.

17.4 Pulse & Remote Support

Pulse	Settings > Pulse — enables anonymous diagnostic data transmission to Nutanix.
Remote Support	Settings > Remote Support — allows Nutanix support to tunnel into cluster for troubleshooting.
Remote Tunnel	Creates a secure SSH tunnel from cluster to Nutanix support infrastructure.

SECTION 18

acli, ncli & Additional Command Line Reference

18. Command Line Interface Reference

18.1 SSH Access

SSH to CVM	ssh nutanix@<CVM-IP> (password: nutanix/4u or configured password)
SSH to AHV Host	From CVM: ssh root@192.168.5.1 (AHV host internal IP)
Cluster VIP SSH	Connects to whichever CVM currently holds the VIP

18.2 acli — Acropolis CLI

VM Operations

acli — VM commands

acli vm.list                                   # List all VMs
acli vm.get <vm-name>                          # Get VM details
acli vm.create <vm-name>                       # Create VM
acli vm.on <vm-name>                           # Power on
acli vm.off <vm-name>                          # Power off (hard)
acli vm.shutdown <vm-name>                     # ACPI shutdown
acli vm.clone <vm-name> clone_name=<new-name>  # Clone a VM
acli vm.delete <vm-name>                       # Delete VM
acli vm.migrate <vm-name> to_host=<host-IP>    # Live migrate VM
acli vm.nic_list <vm-name>                     # List VM NICs
acli vm.disk_list <vm-name>                    # List VM disks
acli vm.update <vm-name> memory_mb=4096        # Update memory
acli vm.update <vm-name> num_vcpus=4           # Update vCPU
acli vm.snapshot.create <vm-name> snap_name=<snap>  # Snapshot
acli vm.snapshot.list <vm-name>                # List snapshots
acli vm.snapshot.revert <vm-name> snap_id=<id> # Revert snapshot

Network & Image Operations (acli)

acli — network & image

acli net.list                                  # List AHV networks
acli net.create <net-name> vlan=<id>           # Create VLAN network
acli net.delete <net-name>                     # Delete network
acli net.add_dhcp_pool <net-name> start=<ip> end=<ip>  # DHCP pool
acli image.list                                # List images
acli image.create <img-name> image_type=kDiskImage source_url=<url>  # Create image
acli image.delete <img-name>                   # Delete image

18.3 ncli — Nutanix CLI

Cluster Information

ncli — cluster & storage

ncli cluster info                              # Cluster details, IPs, version
ncli cluster get-hypervisor-info              # Hypervisor type and version
ncli cluster status                            # Cluster services status
ncli host list                                 # List all nodes
ncli disk list                                 # List all disks
ncli disk get id=<id>                          # Get disk details
ncli container list                            # List all containers
ncli container create name=<n> sp-name=<pool> # Create container
ncli sp list                                   # List storage pools
ncli vdisk list                               # List vDisks
ncli vm list                                  # List VMs
ncli vm get name=<vm>                         # Get VM details
ncli alert list                               # List all alerts
ncli alert acknowledge id=<id>               # Acknowledge alert
ncli alert resolve id=<id>                   # Resolve alert
ncli event list                              # Event log
ncli pd list                                 # List protection domains
ncli pd get name=<pd>                        # Get PD details
ncli pd list-snaps name=<pd>                 # List snapshots
ncli pd restore-vm name=<pd> snap-id=<id> vm-names=<vm>  # Restore VM
ncli pd add-vms name=<pd> vm-names=<vm>      # Add VM to PD

18.4 Bash Commands & Diagnostics

CVM bash — cluster & diagnostics

cluster status                                # Show all service states on all CVMs
cluster start                                 # Start cluster services
cluster stop                                  # Stop cluster services
cluster version                               # Show AOS version
genesis status                                # Check genesis service
allssh genesis status                         # Run command on ALL CVMs
allssh 'cluster status'                       # Cluster status from all nodes
hostssh '<command>'                           # Run on all AHV hosts
svmips                                        # List all CVM IPs in cluster
ncc health_checks run_all                     # Run all NCC checks
df -h                                         # CVM disk usage
free -g                                       # CVM memory usage
top                                           # Real-time CPU/process
ovs-vsctl show                               # OVS bridge config (run via hostssh)
ping -c 4 <target-ip>                         # Connectivity test

18.5 Key Log Locations

CVM log paths

/home/nutanix/data/logs/stargate.INFO          # Stargate I/O logs
/home/nutanix/data/logs/curator.INFO           # Curator scan logs
/home/nutanix/data/logs/cassandra/             # Cassandra DB logs
/home/nutanix/data/logs/zookeeper/             # Zookeeper logs
/home/nutanix/data/logs/ncc-output.log         # Latest NCC output
/home/nutanix/data/logs/genesis.out            # Genesis service log

SECTION 19

Introduction to Prism Central UI

19. Prism Central (PC) UI

19.1 Prism Central Overview

Prism Central is the centralized management plane for multiple Nutanix clusters. It provides a single UI, API, and policy engine for managing hundreds of clusters across data centers and clouds.

PC Deployment	PC is deployed as a VM (or 3-VM scale-out) on one of the managed clusters.
PC Registration	Each PE cluster must be registered with PC: PE > Settings > Prism Central Registration.
PC Scale-Out	Deploy 3 PC VMs for HA — supports larger environments with more performance.
PC Version	PC has its own release cadence — updated independently of AOS via LCM.
PC URL	https://<PC-IP>:9440 (or PC VIP for scale-out)
Prism Pro	License tier for PC — unlocks advanced features: runways, X-Play, advanced analytics.
X-Play (Playbooks)	Automated action engine in PC — trigger actions on alerts, schedule tasks, send notifications.

19.2 PC Navigation Reference

Infrastructure > Clusters	View all registered clusters — health, version, node count.
Infrastructure > Hosts	Cross-cluster host inventory — filter, search, view individual node metrics.
Infrastructure > VMs	Global VM inventory across all clusters — manage, search, filter all VMs.
Compute & Storage > Images	Global image management — publish images across clusters.
Compute & Storage > Categories	Define and manage VM categories for policy targeting.
Policies > Security	Flow microsegmentation policies — security, isolation, quarantine.
Policies > Protection	Leap protection policies — recovery plans for DR.
Activity > Tasks	Cross-cluster task monitor.
Activity > Alerts	Cross-cluster alert console.
Operations > Playbooks	X-Play automated response playbooks.
Operations > Analysis	Cross-cluster performance analytics.
Operations > Planning	Capacity planning and runway projections (Prism Pro).

SECTION 20

Useful Options for Simplified Operations from Prism Central

20. Prism Central — Simplified Operations

20.1 Categories & Policy-Based Management

Category	Key:Value pair applied to VMs. Example: Environment:Production, AppType:WebServer.
System Categories	Pre-defined: AppType, Environment, AppTier — can be extended.
Create Category	PC > Administration > Categories > + New Category → define key and values.
Apply Category	PC > VMs > select VM > Actions > Manage Categories → assign key:value pairs.
Policy Targeting	Security policies, protection policies, and QoS target categories instead of individual VMs.

20.2 Global VM Search & Filter

PC > VMs → use filter bar to search by name, category, power state, cluster, host
Saved Filters: Save common filter combinations for quick access
Bulk Actions: Select multiple VMs across clusters for bulk power on/off, category assignment

20.3 X-Play (Playbooks — Automation)

Trigger	Alert-based (on specific alert), scheduled (cron), or manual.
Actions	Email, Slack webhook, API call, VM power on/off/restart, snapshot, add memory, run script.
Example Playbook	When VM CPU > 95% for 5 minutes → increase memory by 2GB + send email alert.
Use Cases	Auto-remediation, auto-scaling, compliance enforcement, report generation.
No Code	Visual drag-and-drop workflow builder — no programming required.

20.4 Capacity Runway (Prism Pro)

Runway	Prediction of when cluster will run out of storage, CPU, or memory based on growth trends.
Storage Runway	Projects full date based on current growth rate + efficiency savings.
Compute Runway	Projects when CPU or memory will be exhausted.
What-If	Simulate adding nodes or VMs — see how it changes the runway.
Alert	Runway alert triggered when projected full date falls within 90 days (configurable).

20.5 Multi-Cluster Upgrade Orchestration

PC > LCM > Inventory — discover available upgrades for all clusters
Stage upgrades: select clusters and schedule upgrade window
PC applies rolling upgrades cluster by cluster with pre-checks
Track all upgrade tasks from PC Tasks view

20.6 Reports

Custom Reports	PC > Operations > Reports — create custom reports on VMs, storage, hosts.
Scheduled Reports	Auto-generate and email reports on a schedule (daily/weekly/monthly).
Templates	Pre-built report templates: Cluster Efficiency, VM Inventory, Compliance, Alerts Summary.
Export	Reports export as PDF or CSV.

★ LAB

Lab Cheat Sheet — Quick Reference for Hands-On Practice

Lab Cheat Sheet

🔑

Key Access Information

Prism Elementhttps://<VIP>:9440

Prism Centralhttps://<PC-IP>:9440

Default Useradmin

CVM SSHssh nutanix@<CVM-IP>

CVM SSH Passnutanix/4u

AHV Host SSHssh root@192.168.5.1

Foundation UIhttp://<FndVM>:8000

🌐

Key Port Reference

9440Prism UI / REST API (HTTPS)

22SSH (CVM & AHV host)

2049NFS (hypervisor → CVM)

3260iSCSI data path

2100Zookeeper quorum

2888/3888Zookeeper leader election

8000Foundation web UI

80/443Pulse, LCM, external comms

⬆️

LCM Upgrade Safety Order

Step 1NCC (health check plugin)

Step 2Foundation

Step 3AOS (cluster software)

Step 4AHV Hypervisor

Step 5Firmware (BIOS/BMC/HDD)

Step 6Prism Central (separate)

🔍

RF & Cluster Minimums

RF2 min nodes3 nodes

RF3 min nodes5 nodes

RF2 tolerates1 node failure

RF3 tolerates2 node failures

Metadata RFAlways RF3

Quorum needed>50% of nodes

Block AwarenessReplicas across chassis

📋 Daily Operations CLI Quick Reference

Most-used commands

## HEALTH
ncc health_checks run_all              # Full NCC run
cluster status                         # Service status across all CVMs
allssh 'df -h'                         # Disk space all CVMs
allssh 'free -g'                       # RAM usage all CVMs

## VM OPERATIONS (acli)
acli vm.list                           # All VMs
acli vm.on <name>                      # Power on
acli vm.off <name>                     # Force off
acli vm.shutdown <name>                # Graceful shutdown
acli vm.clone <name> clone_name=<new>  # Clone VM
acli vm.migrate <name> to_host=<ip>   # Live migrate

## CLUSTER INFO (ncli)
ncli cluster info                      # Cluster details
ncli host list                         # Node list
ncli disk list                         # Disk inventory
ncli alert list                        # Active alerts
ncli pd list                           # Protection domains
ncli container list                    # Storage containers

✅ Lab Checklist

☐ Deploy Foundation VM → Access http://Foundation-IP:8000
☐ Discovery nodes via IPMI → Assign Host, CVM, and Cluster IPs
☐ Select AOS + hypervisor bundle → Start Foundation → Wait 45–60 min
☐ Login to Prism Element → Verify cluster health (green)
☐ Configure NTP: Settings > NTP Servers
☐ Configure DNS: Settings > Name Servers
☐ Create storage container with RF2 + inline compression
☐ Upload image: Settings > Image Configuration > + Upload Image
☐ Create AHV network (VLAN): Network > Create Network
☐ Create VM from image: VM > Create VM
☐ Verify VM boots, connect via console
☐ Create Protection Domain → Add VM → Set schedule → Configure remote site
☐ Take ad-hoc snapshot → Restore to new VM
☐ Run NCC: Health > Run NCC → review results
☐ Check Analysis charts: select VM → last 1 hour → IOPS + Latency
☐ Register cluster with Prism Central
☐ Create category in PC → Apply to VM → Create Flow security policy

🚨 Troubleshooting Quick Guide

Prism UI unavailable	SSH to CVM → `genesis status` → check if prism service is running → `genesis restart prism`
CVM not starting	SSH to AHV host → `virsh list --all` → `virsh start <cvm-name>`
Disk failure	Prism > Hardware > select disk > Remove Disk → physically replace → re-add via Prism
Node unreachable	Ping IPMI IP → check power → use IPMI KVM to see console → check network cable
Replication failing	Check remote site connectivity → `ncli pd list` → verify CVM reachability at remote site
High CPU on CVM	Check Curator scan: `curator_cli get_last_successful_scan_stats` → check for ongoing rebuild
Storage near full	Check thin provision overage → run Curator dedup/compression → add node or expand
Under-replicated data	Check for node/disk failures → check rebuild progress: `curator_cli master_get_master_location`
VM not starting	Check if host has enough RAM → verify storage container accessible → `acli vm.get <name>`

📚 Full Terminology Glossary

Term	Definition
AOS	Acropolis Operating System — Nutanix cluster software
AHV	Acropolis Hypervisor — Nutanix's KVM-based hypervisor
CVM	Controller VM — manages storage I/O on each node
DSF	Distributed Storage Fabric — pooled node-local storage
RF	Replication Factor — number of data copies (RF2=2, RF3=3)
PE	Prism Element — per-cluster management UI
PC	Prism Central — multi-cluster management UI
LCM	Life Cycle Manager — automated upgrade tool
NCC	Nutanix Cluster Check — automated health check
PD	Protection Domain — DR grouping of VMs
EC-X	Erasure Coding — space optimization for cold data
ILM	Information Lifecycle Management — auto data tiering
Oplog	Write buffer — SSD-backed volatile write cache
Extent	1 MB data chunk — basic I/O unit
Extent Group	4 MB collection of extents — replication/dedup unit
Stargate	CVM service managing data I/O
Curator	CVM service for background maintenance tasks
Cerebro	CVM service for DR replication
Medusa	CVM metadata abstraction service
Cassandra	Distributed metadata database on CVMs
Zookeeper	Distributed coordination and cluster state manager
Genesis	CVM bootstrap and service management
Foundation	Nutanix node imaging and cluster creation tool
Pulse	Optional telemetry/diagnostic data service
FSVM	File Server VM — provides NFS/SMB in Nutanix Files
OVS	Open vSwitch — software virtual switch in AHV
VIP	Virtual IP — floating management IP for cluster
NCP-MCI	Nutanix Certified Professional — Multicloud Infrastructure
RPO	Recovery Point Objective — max acceptable data loss
RTO	Recovery Time Objective — max time to restore service

End of Nutanix Foundations Technical Guide

20 Sections · 200+ Terms · 60+ CLI Commands · Lab Cheat Sheet · AOS · AHV · Prism Element · Prism Central

☁️ Nutanix Foundations Guide

Nutanix Foundations

Introduction & Evolution of Enterprise Infrastructure

Understanding Physical Construct & Basic Behavior

Logical Construct & Detailed Behavior

How Data Resiliency Is Managed by Nutanix

Space Optimization Features & Impact on Data Path

Basic Networking Concepts Related to Nutanix Architecture

Understanding Concepts of Native Disaster Recovery

Nutanix Files, Self-Service, Flow & Leap

Introduction to Foundation & Node Imaging

Portal, Test Drive, University & Certification

Prism Element UI Dashboard & Operations

Hands-On CUD Operations (Create, Update, Delete)

Day 2 Operations

Introduction & Configuration of Data Protection

Health, Analysis, Alerts & Tasks in Prism Element

Showcasing Failures & Impact on Cluster

Explaining & Configuring Options from Settings

acli, ncli & Additional Command Line Reference

Introduction to Prism Central UI

Useful Options for Simplified Operations from Prism Central

Lab Cheat Sheet — Quick Reference for Hands-On Practice

Key Access Information

Key Port Reference

LCM Upgrade Safety Order

RF & Cluster Minimums