NCA: Nutanix Certified Associate Study Guide
The Nutanix Certified Associate (NCA) validates foundational knowledge of Nutanix hyperconverged infrastructure, covering core architecture, cluster management with Prism, VM and storage administration, AHV networking, data protection, and monitoring. It is aimed at IT professionals new to Nutanix - administrators, support staff, and sales engineers - who need to demonstrate baseline competency operating an AOS cluster. The 90-minute exam has roughly 60 questions, a passing score of 600 (on a scaled basis), and focuses on operating the platform through Prism rather than deep design or troubleshooting.
Domain 1: Nutanix Concepts and Architecture
- Hyperconverged infrastructure (HCI) collapses compute, storage, and virtualization into a single software-defined layer running on commodity x86 servers, eliminating the separate SAN/NAS array and the dedicated storage network fabric.
- The legacy 3-tier model separates compute servers, a centralized storage array (SAN or NAS), and the storage network (e.g. Fibre Channel switches), with each tier procured, scaled, and administered independently - often by different teams.
- A Nutanix node is a single physical x86 server; a cluster is a logical group of nodes that work together as one system. A block is the chassis that houses one or more nodes.
- Every node runs a hypervisor (AHV, ESXi, or Hyper-V) for user VMs plus a Controller VM (CVM) that provides AOS storage services; together they form the HCI node.
- AOS (Acropolis Operating System) is the core software that runs in the CVMs and delivers the Distributed Storage Fabric plus data services such as snapshots, replication, deduplication, and compression. AOS is hypervisor-agnostic.
- The Distributed Storage Fabric (DSF) aggregates the local SSDs and HDDs from every node into a single, shared, cluster-wide storage pool that any VM on any node can access.
- Data redundancy is achieved by keeping redundant copies of data on different nodes using a Replication Factor: RF2 keeps two copies (tolerates one component failure), RF3 keeps three copies (tolerates two simultaneous failures).
- Data locality means a VM's I/O is served by the CVM on the same node where the VM runs; frequently accessed (hot) data is migrated to the local node for low-latency access.
- Tiering serves hot, frequently accessed data from fast SSD while colder data is moved down to HDD, optimizing both performance and capacity in hybrid clusters.
- Self-healing: when a node or disk fails, the cluster automatically rebuilds the missing data copies onto other healthy nodes to restore the configured redundancy without manual intervention.
- Nutanix scales out horizontally - capacity and performance grow by adding nodes one at a time, and new resources join the cluster automatically (web-scale architecture).
- A key advantage of HCI over 3-tier is that compute and storage scale together incrementally, avoiding the large, lumpy storage-array purchases that 3-tier often forces.
- If a host fails, HA automatically restarts the affected user VMs on other healthy nodes in the cluster.
- Hybrid/multicloud consistency: Nutanix runs the same platform and management experience on-premises and in public cloud (e.g. Nutanix Cloud Clusters / NC2).
Domain 2: Managing the Cluster with Prism
- Prism Element (PE) is the built-in, per-cluster management interface; it can fully manage a single cluster on its own, and Prism Central is optional.
- Prism Central (PC) is a separate, deployable appliance that provides single-pane-of-glass management and cross-cluster monitoring across multiple registered clusters.
- Registering a cluster with Prism Central establishes a management connection so PC can monitor and manage it; unregistering severs that link. If the connection is impaired, the cluster keeps running but PC can no longer manage it.
- The Home (main) dashboard in PE is the default landing page and shows consolidated widgets: Hypervisor Summary, Cluster-wide Controller IOPS, CPU/Memory usage, Storage Summary, Health, and recent Alerts and Events.
- The Cluster-wide Controller IOPS widget reports the combined storage IOPS served by all CVMs, reflecting the real storage workload because every VM's I/O is serviced by the DSF.
- The Hardware view manages the physical layer: a Diagram tab shows blocks and the nodes within them, while the Table tab lists hosts (CPU, memory, model) and disks (capacity, tier, status).
- The Storage view handles logical storage - it shows the storage pool, storage containers, total vs used capacity, and efficiency/data-reduction ratios.
- The VM view lists every VM with per-VM CPU, memory, IOPS, and power state, and provides lifecycle actions (power on/off, update, clone, launch console).
- Launching the console from a selected VM opens a remote console session to that VM's guest operating system.
- The Health widget/dashboard aggregates NCC health-check results across VMs, hosts, and disks, color-coded green (Good), yellow (Warning), red (Critical).
- Cluster-wide settings (name, NTP, DNS, SMTP, authentication, licensing) are configured under the Settings (gear) menu in Prism.
- Cluster details such as the cluster name, virtual IP, and configuration are accessed from the cluster name menu at the top of Prism Element.
- The view-selector dropdown next to the Home title switches between dashboards (Home, VM, Storage, Hardware, Network, Data Protection, Analysis, Alerts, Health).
- Categories in Prism Central are key-value pairs used to logically tag entities (such as VMs) so that policies and operations can target them as a group.
Domain 3: Virtual Machine Management
- In the AHV Create VM dialog, the vCPU(s) field sets the number of virtual sockets and Cores per vCPU sets cores each, so total logical processors = vCPUs x cores per vCPU.
- The Memory field is expressed in GiB by default; the value entered is what the guest OS sees as installed RAM.
- Adding a disk with the Allocate on Storage Container option provisions a new empty, thin-provisioned vDisk whose capacity is carved from the selected storage container (space is consumed only as data is written).
- ISO and disk images must first be uploaded to the Image Service (Image Configuration) before they can be attached to a VM.
- To install an OS, attach a disk allocated on a storage container plus a CD-ROM with the installation ISO (via Clone from Image Service); set the boot device order accordingly, then power on.
- After OS installation, eject (unmount) the ISO from the CD-ROM so the VM boots from its disk instead of the installation media.
- A VM's NIC must be attached to a network (a VLAN-backed subnet) to pass guest traffic; attaching it to a managed network with IPAM enabled lets the VM receive an IP automatically.
- Cloning copies a fully configured source VM into one or more new VMs - the fast way to deploy many identical VMs (e.g. clone a template VM ten times).
- Snapshots capture a point-in-time state; you can restore a VM from a snapshot or deploy a new VM from an existing snapshot.
- Changing vCPU count typically requires the VM to be powered off (or power-cycled): power off, update vCPUs, power back on.
- Memory hot-add may fail if no single host has enough free physical RAM for the new total; otherwise the guest must support hot-add or the VM must be power-cycled.
- Update VM modifies an existing VM's configuration - vCPUs, memory, disks, and network adapters - through the Update VM dialog (disks under the Disks section).
- Power operations include Power On/Off, plus Guest Shutdown and Guest Reboot, which gracefully signal the guest OS (these graceful options require guest integration).
- Nutanix Guest Tools (NGT) is the in-guest software package that enables features like graceful shutdown, application-consistent snapshots, self-service restore, and VM mobility.
Domain 4: Storage Management
- A storage pool is a logical grouping of all physical storage devices (SSDs and HDDs) across all nodes; a default storage pool containing all cluster disks is created automatically at cluster creation.
- Best practice is to use a single storage pool, because pooling all disks maximizes flexibility and is the recommended default.
- The storage hierarchy is: physical drives -> storage pool -> storage container(s) -> VM vDisks. Containers are logical subdivisions of the pool.
- A storage container is the object where data-reduction and data-protection settings - compression, deduplication, erasure coding, and replication factor - are configured. Creating one requires a name and the pool it is carved from.
- On AHV, a new container is automatically presented and mounted to the hosts as an NFS datastore (on Hyper-V it is presented as an SMB share); no manual mount is needed.
- To apply different data-reduction policies to different workloads, create separate containers (e.g. one with compression for an archive, one without for latency-sensitive SQL).
- Advertised capacity is an optional logical quota on a container that caps how much capacity it can report/consume, reserving space for other containers; without it, a container can grow to fill the entire pool.
- Reserved capacity (capacity reservation) guarantees a minimum amount of pool space for a container so it is always available to it.
- Advertised capacity is a logical quota, whereas raw capacity is the total physical disk space in the pool before data reduction and replication overhead.
- Compression reduces data footprint and gives the greatest savings on cold data that is seldom rewritten; it can be inline or post-process.
- Deduplication eliminates redundant copies by generating a fingerprint (hash) for each data block and matching identical fingerprints, storing only one copy.
- Erasure coding (EC-X) reduces the capacity overhead of replication for cold data by computing parity strips instead of full copies.
- Thin provisioning means a vDisk consumes physical space only as data is written - e.g. ~200 GB written consumes roughly 200 GB regardless of the disk's advertised size.
- VM vDisk data is distributed across the physical drives of the storage pool and presented through the container the disk was assigned to.
Domain 5: Networking and Data Protection
- On AHV, the default virtual switch vs0 is created automatically and maps to the default Open vSwitch bridge br0 on every host, providing Layer 2 connectivity for VM NICs and the CVM.
- AHV networking is implemented with Open vSwitch (OVS); virtual switches are managed centrally in Prism so configuration applies consistently across all AHV hosts.
- A subnet/network in AHV is created by assigning a VLAN ID (e.g. VLAN ID 30); attaching a VM NIC to it places that VM's traffic on the corresponding VLAN.
- A managed network has AHV-provided IP addressing (IPAM) - configure an IP address pool/range, a default gateway, and the network/prefix - so AHV assigns guest IPs without an external DHCP server.
- An unmanaged network defines only the VLAN ID; VMs must get IPs from an external DHCP server on that VLAN or be assigned static IPs in the guest.
- The physical NICs connecting a host to upstream switches are grouped into an uplink bond for redundancy; supported load-balancing modes include active-backup, balance-slb, and balance-tcp (LACP).
- Upstream switch ports for AHV uplinks should be configured as trunk ports that allow the required VLANs; if a trunk does not allow the VM's VLAN, the VM loses connectivity.
- Traffic on the native VLAN of the uplink is sent untagged; a VM placed on the native VLAN sends untagged frames.
- CVM management and storage backplane traffic can be isolated from VM traffic using network segmentation for security and performance.
- A Protection Domain (in PE Data Protection) groups entities (VMs/volume groups) that are snapshotted and replicated together on a common schedule; the schedule defines recurrence interval and retention.
- A consistency group within a Protection Domain ensures that multiple related VMs are snapshotted at the same instant - place interdependent VMs in the same consistency group.
- Crash-consistent snapshots capture the VM's disk state as if power was lost (no guest quiescing); application-consistent snapshots quiesce the guest via NGT for transactionally clean data.
- Snapshots are created near-instantly using redirect-on-write with negligible performance impact; when retention is reached the oldest snapshot is automatically expired (e.g. keeping only the 5 most recent).
- Nutanix Leap (Protection Policies / Recovery Plans in Prism Central) provides orchestrated DR and runbook-style failover across sites, beyond the per-cluster Protection Domains in PE.
Domain 6: Monitoring, Health, and Maintenance
- Alerts are notifications of conditions that may need attention; events are a chronological record of actions or state changes that have occurred in the cluster.
- Alert severities are Info, Warning, and Critical. Critical (red) indicates a problem that can impact availability, data resiliency, or performance (e.g. failed disk, unreachable node, pool nearly full).
- Acknowledging an alert marks it as seen for other admins but does not resolve the underlying condition; resolving/closing it reflects the issue being addressed.
- An alert message includes the title, severity, affected entity, time created, and suggested resolution steps; the Alerts dashboard groups alerts by severity.
- The Health dashboard provides a consolidated, color-coded view (Good/Warning/Critical) of hosts, disks, storage pools, containers, and VMs, driven by background NCC checks.
- NCC (Nutanix Cluster Check) is the built-in health-diagnostic framework included with every cluster; it verifies configuration and stability of cluster components.
- Run all checks with the command 'ncc health_checks run_all'; run a specific check with 'ncc health_checks <category> <check_name>'.
- NCC checks return results of PASS (healthy / meets criteria), WARN (a potential issue to review, not yet a hard failure), FAIL (the tested condition is unhealthy), and ERR (the check itself could not complete).
- NCC supports scheduling so health checks run automatically at defined intervals, and output can be collected as a log bundle for Nutanix Support.
- The Analysis page creates and overlays performance charts (CPU, memory, IOPS, latency, throughput) for a VM, host, or the whole cluster over a selected time range, with multiple metrics/entities per chart.
- The Data Resiliency Status widget shows whether the cluster currently has enough redundancy to tolerate a disk or node failure and still rebuild a redundant copy (OK in green when it can).
- The Capacity Runway widget projects how long until resources (storage, CPU, memory) are exhausted, helping plan capacity or workload proactively from usage trends.
- Pulse (configured under cluster Settings) sends cluster diagnostic data to Nutanix so issues can be proactively detected and support cases opened before they impact the customer.
- Alert email notifications are configured in Alert Email Configuration under Prism settings, where recipients and SMTP delivery are defined.
NCA exam tips
- Know the architecture cold: node vs cluster vs block, what the CVM does, what AOS is, and where the Distributed Storage Fabric fits. Many questions test these definitions and the HCI-vs-3-tier contrast directly.
- Distinguish Prism Element (single cluster, always present) from Prism Central (optional, multi-cluster single pane of glass), and know which dashboard/widget shows what (Hardware = physical, Storage = logical, Analysis = charts, Health = NCC results).
- Memorize the storage hierarchy - physical disks -> storage pool -> container -> vDisk - and remember that data-reduction and replication-factor settings live at the container level, with a single pool being the recommended default.
- For AHV networking, be precise about managed (IPAM provides IPs) vs unmanaged (external/static IPs) networks, that vs0 maps to br0, and that uplink trunk ports must allow the VM's VLAN or connectivity breaks.
- Learn the NCC command syntax and result types (PASS/WARN/FAIL/ERR), the difference between alerts and events, and snapshot consistency types (crash-consistent vs application-consistent via NGT). These are common, easily-scored points.
Study guide FAQ
How many questions are on the NCA and how long do I have?
The NCA is a 90-minute exam of roughly 60 multiple-choice questions. The passing score is 600 on Nutanix's scaled scoring model, so aim to be comfortable across all six domains rather than mastering just one.
Do I need hands-on cluster experience to pass?
It helps but is not strictly required - the NCA is an associate-level, foundational exam focused on concepts and operating the platform through Prism. Working through the free Nutanix Test Drive / hosted clusters and the official NCA learning path to click through Prism, create a VM, and run an NCC check will solidify the material quickly.
Which hypervisor does the exam focus on?
AHV, Nutanix's built-in hypervisor based on KVM, is the primary focus - especially for VM creation, Open vSwitch/vs0 networking, and IPAM. Be aware AOS also supports ESXi and Hyper-V (which mounts containers as SMB shares rather than NFS), but most procedural questions assume AHV.
What is the difference between Prism Element and Prism Central for the exam?
Prism Element is the management UI built into every cluster and can fully manage that one cluster by itself. Prism Central is an optional, separately deployed appliance that registers multiple clusters for single-pane-of-glass monitoring and adds features like categories, Leap/DR orchestration, and cross-cluster reporting.