Nutanix AHV and Prism Architecture

3 Nov 2026, 18:2115 min readby ITYantra Admin

Comprehensive guide to Nutanix AHV hypervisor and Prism management architecture including distributed storage, AHVM, and cluster operations.

# Nutanix AHV and Prism Architecture

Introduction

Nutanix AHV (Acropolis Hypervisor) is a KVM-based enterprise hypervisor integrated into the Nutanix hyperconverged infrastructure (HCI) platform. AHV is managed through Prism, Nutanix's centralized management interface. Unlike traditional virtualization platforms that require separate hypervisor, storage, and management layers, Nutanix combines compute and storage into a single scalable building block called a node.

This tutorial covers Nutanix AHV architecture, Prism management, distributed storage (Distributed Storage Fabric), AHV virtual machine operations, cluster administration, and hyperconverged infrastructure best practices.

By the end of this guide, you will understand Nutanix HCI architecture, how AHV and Prism work together, and how to administer a Nutanix cluster.

Technical Background

Nutanix HCI Architecture

Nutanix hyperconverged infrastructure combines:

  • **Compute**: x86 servers running AHV hypervisor
  • **Storage**: Direct-attached storage aggregated into distributed storage fabric
  • **Management**: Prism Central and Prism Element for centralized management

AHV (Acropolis Hypervisor)

AHV is a type-1 hypervisor based on KVM (Kernel-based Virtual Machine), customized by Nutanix for HCI workloads. Key components:

  • **AHVM (AHV Manager)**: Management agent running on each node
  • **CVM (Controller Virtual Machine)**: Storage services VM running on each node, handles distributed storage operations
  • **Stargate**: Storage API gateway running in CVM
  • **Curator**: Distributed storage manager
  • **Zeus**: Block storage service

Prism

Prism is Nutanix's management interface:

  • **Prism Element**: Built into each cluster, manages individual cluster
  • **Prism Central**: Centralized management across multiple clusters

Distributed Storage Fabric (DSF)

DSF aggregates local SSDs/HDDs from all nodes into a single distributed storage pool with:

  • Data redundancy (RF2, RF3)
  • Erasure coding
  • Inline compression and deduplication
  • Self-healing

Prerequisites

  • Nutanix hardware nodes (certified)
  • Network connectivity between nodes
  • Prism Element access (IPMI or dedicated management)
  • Nutanix licenses
  • Understanding of HCI concepts

Architecture

Nutanix Cluster Architecture

graph TB
    subgraph Node_1 [Node 1]
        AHV_1[AHV Hypervisor]
        VM_1[VMs]
        CVM_1[CVM - Controller VM]
        SSD_1[(SSD/HDD)]
    end
    subgraph Node_2 [Node 2]
        AHV_2[AHV Hypervisor]
        VM_2[VMs]
        CVM_2[CVM - Controller VM]
        SSD_2[(SSD/HDD)]
    end
    subgraph Node_3 [Node 3]
        AHV_3[AHV Hypervisor]
        VM_3[VMs]
        CVM_3[CVM - Controller VM]
        SSD_3[(SSD/HDD)]
    end
    
    subgraph Prism [Prism Element / Central]
        PC[Prism Management]
    end
    
    AHV_1 <==> AHV_2 <==> AHV_3
    CVM_1 <==> CVM_2 <==> CVM_3
    
    PC --> AHV_1
    PC --> AHV_2
    PC --> AHV_3
    
    VM_1 --> SSD_1
    VM_2 --> SSD_2
    VM_3 --> SSD_3

Storage Path

VM Write ──→ AHV ──→ CVM ──→ Distributed Storage (all nodes)
VM Read  ──→ AHV ──→ CVM ──→ Local or Remote Storage

Step-by-Step Implementation

Step 1: Cluster Discovery and Setup

**Purpose**: Discover Nutanix nodes and initialize the cluster.

**Procedure**:

1. Connect all nodes to network 2. Power on nodes 3. Access one node via IPMI or laptop connected to node 4. Run Foundation (Nutanix cluster deployment tool) from laptop:

   ./foundation -m <node-ips> -u admin -p <password> -i <image>

5. Foundation images nodes with AHV and creates cluster 6. Access Prism Element at cluster VIP

**Commands**: None (Foundation tool GUI/CLI)

**Expected output**: Nutanix cluster created; Prism Element accessible.

**Verification**:

  • Prism Element URL accessible
  • All nodes shown in cluster health dashboard

Step 2: Configure Prism Element

**Purpose**: Initial cluster configuration through Prism Element.

**Procedure** (Prism Element web interface):

1. Login to Prism Element (https://cluster-vip:9440) 2. Configure cluster name 3. Configure NTP, DNS, SMTP 4. Configure storage containers 5. Activate licenses 6. Create admin users

**Expected output**: Cluster configured and operational.

**Verification**:

  • Prism dashboard shows green health
  • Storage containers configured
  • Licenses active

Step 3: Configure Storage Container

**Purpose**: Create storage containers for VM disks.

**Procedure** (Prism Element):

1. Navigate to Storage → Storage Containers 2. Click + Create Storage Container 3. Enter container name 4. Select storage pool 5. Configure compression/deduplication 6. Set ad-hoc compression 7. Save

**Commands** (via Nutanix CLI - ncli):

# List storage containers
ncli container list

# Create storage container
ncli container create name=vm-container \
  storage-pool=Default-Store \
  compression-enabled=true \
  deduplication-enabled=true

**Expected output**: Storage container created and available for VM disks.

**Verification**:

  • Prism shows storage container with capacity
  • Container available when creating VMs

Step 4: Create Virtual Machine

**Purpose**: Deploy a VM on AHV.

**Procedure** (Prism Element):

1. Navigate to VMs → Create VM 2. Enter VM name and description 3. Select vCPU count and memory 4. Select storage container 5. Add network adapter (select subnet) 6. Add disk (select size and container) 7. Add ISO for OS installation (optional) 8. Save and power on

**Commands** (via ncli):

# Create VM
ncli vm create name=web-server-01 \
  num-vcpus=2 memory=4G \
  container=vm-container \
  network=primary-network \
  disk-size=40G

# Start VM
ncli vm start uuid=<vm-uuid>

**Expected output**: VM created and running on AHV.

**Verification**:

  • Prism shows VM power state: On
  • VM accessible via console or network

Step 5: Configure High Availability

**Purpose**: Enable AHV HA for automatic VM restart on host failure.

**Procedure** (Prism Element):

1. Navigate to Cluster → High Availability 2. Enable AHV HA 3. Configure failover tolerance (default: 1 host) 4. Set VM VM monitoring (optional) 5. Save

**Explanation**: AHV HA monitors hosts and restarts VMs on remaining hosts if a host fails. Uses Cassandra-based heartbeating between CVMs.

**Expected output**: HA enabled; VMs protected against host failure.

**Verification**:

  • Prism shows HA status: Enabled
  • Simulate host failure; VMs restart on remaining hosts

Step 6: Cluster Health Monitoring

**Purpose**: Monitor cluster health and performance.

**Procedure** (Prism Element):

1. Dashboard → Health: Check cluster, storage, and VM health 2. Dashboard → Performance: View CPU, memory, storage, network metrics 3. Alerts → View active alerts 4. Health → Run health checks

**Commands** (via ncli):

# Check cluster health
ncli cluster health

# Check storage status
ncli storagepool list

# Check host status
ncli host list

**Expected output**: All health checks pass; no active alerts.

**Verification**:

  • Prism dashboard shows all green
  • ncli cluster health returns healthy

Configuration Examples

Prism Element URLs

Prism Element: https://<cluster-vip>:9440
Prism Central: https://<prism-central-ip>:9440

Nutanix CLI (ncli) Examples

# Cluster info
ncli cluster get-info

# Storage pool list
ncli storagepool list

# Host list
ncli host list

# VM list
ncli vm list

# Network list
ncli network list

AHV Host Commands (via SSH to CVM)

# Check AHV host status
virsh list --all

# Check storage pools
virsh pool-list

# Check cluster status
cluster health

Storage Container Configuration

Container: vm-container
Storage pool: Default-Store
Compression: Enabled
Deduplication: Enabled
Ad-hoc compression: Default
Cloning: Enabled

Verification

Cluster Health Verification

ncli cluster health
# Expected: Cluster health: OK
# All nodes healthy
# All CVMs healthy
# Storage healthy

VM Operations Verification

ncli vm list
# Shows all VMs with name, UUID, power state, host

ncli vm get-uuid name=web-server-01
# Returns VM UUID for further operations

Storage Verification

ncli storagepool list
# Shows storage pool capacity, used, free

ncli container list
# Shows storage containers with compression/dedup status

Expected Results

  • Cluster health: OK
  • All nodes and CVMs healthy
  • Storage pools online with adequate capacity
  • VMs running with proper power state
  • No active alerts

Troubleshooting

| Symptom | Possible Cause | Diagnostic Command | Corrective Action | Verification | |---------|---------------|-------------------|-------------------|--------------| | Cluster health degraded | Node or CVM down | `ncli cluster health` | Fix node/CVM issue; check network | Health returns OK | | VM won't start | Insufficient resources | `ncli vm list` | Add host or free resources | VM starts | | Storage full | Too much data | `ncli storagepool list` | Add storage; delete unused VMs | Space available | | Prism not accessible | Management network issue | Ping cluster VIP | Fix network; check Prism services | Prism accessible | | CVM down | CVM crash or host issue | `ncli host list` | Restart CVM or host | CVM healthy | | Slow storage performance | Disk contention or full | Prism performance dashboard | Add SSD tier; rebalance | Performance improves | | HA not working | HA not configured | Prism HA settings | Enable AHV HA | HA protects VMs | | Node not in cluster | Network or image issue | Foundation logs | Re-image node with Foundation | Node joins cluster |

Best Practices

1. **Use RF2 (replication factor 2) minimum**: Protects against single node failure. 2. **Size cluster for N+1 failure**: Cluster should operate with one host down. 3. **Use Prism Central for multi-cluster**: Centralized management across multiple Nutanix clusters. 4. **Enable compression and deduplication**: Reduces storage consumption with minimal performance impact. 5. **Monitor CVM health**: CVMs are critical for storage operations. 6. **Use AHV HA**: Automatic VM restart on host failure. 7. **Keep Nutanix software updated**: Apply LCM (Lifecycle Manager) updates regularly. 8. **Document cluster design**: Record RF, storage containers, networks, and host IPs. 9. **Test failover regularly**: Verify HA works during maintenance windows. 10. **Use consistent node models**: Avoid mixing different node generations in same cluster.

Conclusion

Nutanix AHV and Prism deliver hyperconverged infrastructure with integrated compute, storage, and management. AHV provides KVM-based virtualization optimized for HCI, while Prism Element and Central provide centralized management across clusters. The distributed storage fabric, CVM architecture, and AHV HA create a resilient, scalable platform for enterprise virtualization. Understanding AHV components, Prism management, and cluster operations is essential for administering Nutanix environments effectively.

References

  • Nutanix Documentation Portal (https://portal.nutanix.com/)
  • Nutanix AHV Administration Guide (https://portal.nutanix.com/)
  • Nutanix Prism Element Guide (https://portal.nutanix.com/)
  • Nutanix Bible (https://nutanixbible.com/)
  • Nutanix CLI Reference (ncli) (https://portal.nutanix.com/)
#nutanix#ahv#prism#hypervisor#virtualization#hci#distributed-storage#aosp