Nutanix NCA common mistakes
Nutanix NCA · Understand the traps and how to avoid them.
Answering 'log into Prism Central' for tasks that are single-cluster operations (creating a storage container, adding a node, configuring a VLAN-backed network), and answering 'Prism Element' for multi-cluster or policy tasks (categories, IAM, Flow, Calm).
Both consoles share the same look and many of the same menus, so candidates who have only clicked through one of them assume the other does the same job.
What to do instead
Build a two-column scope table and drill it: Prism Element = per-cluster configuration and monitoring (storage containers, networks, cluster expansion, alerts for that cluster); Prism Central = multi-cluster inventory, categories, RBAC/IAM, Flow microsegmentation, Calm, and cross-cluster reporting.
Treating redundancy factor as a performance or licensing setting and getting capacity/RPO questions wrong - e.g. saying an RF2 cluster survives two simultaneous node failures, or that usable capacity equals raw capacity.
RF is presented as a simple dropdown, so candidates memorize the label without internalizing that RF2 keeps two copies (one failure tolerated, roughly half the raw capacity usable) and RF3 keeps three copies (two failures tolerated).
What to do instead
Memorize the pairs: RF2 = 2 copies, tolerates 1 failure, ~50% usable; RF3 = 3 copies, tolerates 2 failures, ~33% usable. Then re-read any capacity or failure-tolerance question and state which copy count it is really asking about.
Using 'storage pool', 'container', 'vDisk' and 'extent' interchangeably in answers about where RF, compression, or advertised capacity is actually configured.
The terms all appear on the same Prism storage screens, and no single screen labels the full hierarchy, so recall blurs them together.
What to do instead
Draw the hierarchy once from memory: physical disks are grouped into a storage pool; a container is the logical layer that holds the RF setting, dedupe/compression and advertised capacity; VMs write vDisks, which are broken into extents on disk. Answer by locating which layer the question names.
Picking the wrong data-protection feature - describing Metro Availability as scheduled replication, or Async DR as zero-RPO - when the question describes a specific RPO or site topology.
All three appear under the same Data Protection area of Prism and all involve replicating to a second site, so candidates match on 'DR' rather than on the RPO and topology wording.
What to do instead
Anchor each feature to one keyword: Metro Availability = stretched cluster, synchronous, zero RPO, both sites active; NearSync = RPO of minutes via frequent scheduled snapshots; Async DR = scheduled replication with an RPO measured in hours/schedule interval. Scan the question for 'synchronous', 'stretched', or a schedule.
Assuming AHV needs vCenter or VMware port groups, or that virtual networking is configured on the hypervisor host rather than as a Prism-managed network object.
Candidates with a VMware background carry over vSwitch/port-group mental models and answer from habit instead of from what AHV actually does.
What to do instead
Fix the model: AHV is Nutanix's native hypervisor managed by Prism, uses its own virtual switch with no vCenter dependency, and networks are created as Prism network objects with a VLAN ID and then attached to VMs. Practise the create-network-then-attach-VM-nIC sequence.
Defaulting to the most powerful role when a question asks for least privilege - e.g. granting cluster admin to a helpdesk user who only needs to view alerts and VM status.
The Admin role is the only one most candidates have used, and the finer-grained roles and role mapping to directory groups are easy to skip during study.
What to do instead
Learn the role ladder and what each unlocks: view-only/monitoring roles for read access to health and alerts, VM-admin type roles for VM operations without cluster configuration, cluster admin for everything. Also remember directory (AD/LDAP) integration with role mapping to groups, done at Prism Central for multi-cluster environments.
Relying only on Prism dashboard alerts for cluster health questions and ignoring NCC health checks and Pulse telemetry, or mixing up what each one does.
All three live in the same Health/Monitoring area and all sound like 'the cluster checks itself', so the distinct job of each gets flattened.
What to do instead
Separate them: Prism alerts and alert policies notify on events/thresholds; NCC runs the on-demand/periodic cluster health check suite and is what you run before upgrades or when diagnosing; Pulse is the automated telemetry that sends health and configuration data to Nutanix support.
Answering that cluster expansion or a fresh install is done from Prism Central, or that adding a node requires re-imaging the whole cluster.
Prism Central as the 'single pane' is emphasized so heavily that candidates assume every lifecycle action starts there, and they over-estimate how invasive adding a node is.
What to do instead
Sequence it explicitly: image the new node with Foundation, rack and cable it, then use Prism Element's expand-cluster workflow on the existing cluster to add the node; data is redistributed without rebuilding the cluster. Prism Central then just picks up the expanded cluster.
Skipping compatibility checks and assuming AOS, AHV, Prism Central and firmware can be upgraded in any order or all through one tool.
Upgrades look like a single 'update' button, and the compatibility matrix and LCM inventory step are easy to skim past during study.
What to do instead
Remember the discipline: run an LCM inventory first, check the compatibility/interoperability matrix for the target AOS and AHV versions, upgrade Prism Central before the clusters it manages, and remember LCM is the tool for firmware and software baseline updates.
Treating Nutanix snapshots as full independent copies or as a substitute for replication, and therefore answering 'snapshot' for backup/DR scenarios.
Snapshots are described as space-efficient and instant, which leads candidates to assume they are self-contained backups or that they leave the cluster.
What to do instead
State both properties every time: Nutanix snapshots are metadata-based and space-efficient, and they stay on the same cluster; off-cluster protection requires replication via protection domains/async DR or an external backup target. If the question mentions another site or media failure, snapshots alone are not the answer.