NPS key terms

A service virtual machine that runs on every Nutanix node and handles storage I/O, cluster communication, and data services for that host. Guest VMs reach cluster storage through their local CVM.

The CVM service that owns the storage stack on a node. It receives I/O from the hypervisor, manages vDisks and extents, and takes part in data replication across the cluster.

A process inside each controller VM that starts and supervises other Nutanix services, runs local health checks, and helps coordinate cluster-wide state such as which node holds a service role.

A distributed database running in the controller VMs that stores cluster metadata such as vDisk locations, configuration settings, and health information. Its data is replicated so any node can locate stored blocks.

A coordination service used by Nutanix software to elect leaders and hold small amounts of shared cluster state, so services on different nodes agree on which node owns a role.

The Nutanix software layer installed on every node of a cluster. It provides the distributed storage fabric, virtualization management, and the cluster services that tie the nodes together.

The software-defined storage layer that pools local disks from all nodes into one shared datastore. It decides where data lives, keeps copies in sync, and routes I/O so VMs see a single storage fabric.

A logical grouping of the physical drives (SSD, HDD, or NVMe) across cluster nodes. Storage containers are carved out of a pool, and the pool defines the raw capacity the cluster can use.

A logical partition inside a storage pool that VMs and datastores consume. It carries the policies applied to its data, such as replication factor, compression, deduplication, and advertised capacity.

A virtual disk presented to a guest VM. It is divided into extents spread across the cluster, and its metadata maps logical blocks to physical locations so the guest sees an ordinary disk.

A fixed-size chunk of disk data, typically one megabyte, that is the basic unit of placement and replication across controller VMs. Many extent groups together make up a vDisk.

The number of copies of data the cluster maintains. A factor of two keeps two copies on separate nodes; a factor of three keeps three, tolerating more failures at the cost of extra capacity.

A capacity-saving method that stores data plus parity fragments across nodes instead of full duplicate copies. It uses less space than replication while still surviving the loss of a node.

A named collection of VMs, vDisks, or consistency groups grouped for snapshot and replication scheduling. It is the unit that data protection policies act on for local recovery and disaster recovery.

The condition where a VM's read requests are served by the controller VM on the same host, avoiding a network hop. The cluster shifts data toward the hosts that use it most.

Capacity-saving features that remove repeated data blocks and shrink stored data. They are enabled per storage container and may run inline as writes arrive or as later background passes.

The background process that recreates lost data copies on the surviving nodes after a disk or node fails, bringing the container back to its replication factor. It consumes cluster network and disk resources.

A point-in-time, space-efficient copy of a VM's disks or a protection domain's data. Snapshots support rollback, cloning, and as a source for replication to a remote cluster.

The software switch, named vs0, inside the AHV hypervisor that connects VM network interfaces to host uplinks and bridges. It moves traffic between VMs, the host, and the physical network.

The AHV bridge interface that carries the host management address and links the virtual switch to the physical network. Administrators use its address to reach the host and its controller VM.

Two or more physical network cards combined into one logical uplink to provide redundancy or higher throughput. AHV supports modes such as active-backup and link aggregation control protocol.

A logical network segment identified by a tag between 1 and 4094 that separates traffic on shared physical links. In AHV, VLAN tags are applied to VM interfaces and switch uplinks.

The VLAN whose traffic crosses a trunk port without a tag. VM interfaces configured with no VLAN tag use it, so a mismatch between host and switch settings commonly breaks connectivity.

Acropolis Hypervisor, the native KVM-based hypervisor included with the Nutanix software stack. It is administered through Prism and needs no separate hypervisor license or third-party product.

The Nutanix service that handles VM creation, placement, scheduling, and lifecycle operations. It sits between Prism and the hypervisor, giving a common management layer for virtual machines.

A saved virtual machine configuration, optionally including its disks, used to create new VMs with consistent settings. Templates reduce manual setup and are managed through Prism or the image service.

A copy of an existing virtual machine made from a snapshot. It shares unchanged data blocks at first and stores only new writes, so clones appear quickly and use little extra space.

A rule that limits which hosts a VM or group of VMs may run on. It keeps related workloads together or apart to meet performance, licensing, or availability requirements.

Software installed inside a guest operating system that unlocks features such as self-service restore, file-level recovery, and quiesced snapshots. It runs as a small agent within the VM.

A widely used method for passing first-boot configuration such as hostname, users, and scripts to a virtual machine. AHV exposes it to guests through a dedicated initialization device.

The web console and API used to manage a single Nutanix cluster, covering storage, VMs, hosts, alerts, and health. Every cluster has its own instance at the cluster virtual IP.

A separate management appliance that oversees many clusters from one place and adds global capabilities such as categories, projects, role-based access, and cross-cluster dashboards.

Key and value labels attached to VMs, hosts, or images that drive policy, reporting, and automation. Prism Central features such as Flow and Projects rely on them to select objects.

A Prism Central construct that groups virtual machines, users, quotas, and policies together, letting a team manage its own resources within limits set by the administrator.

The Prism permission model that assigns roles, which are sets of privileges, to users or groups and scopes them to objects such as clusters, projects, or categories.

The Prism Central network security feature that applies microsegmentation rules between workloads. Policies are written against categories and enforced at the virtual switch rather than on physical firewalls.

A built-in diagnostic tool that runs a library of health checks against the cluster and reports findings with severity ratings and suggested remediation steps.

The command-line interface available from any controller VM for inspecting and changing cluster settings, hosts, and storage. It is a standard entry point when troubleshooting or scripting tasks.

The automated support feature that periodically sends health and configuration data from the cluster to Nutanix. It enables proactive case creation and faster diagnosis by support engineers.

The Prism feature that inventories the cluster, checks for available software and firmware updates, and applies them in a guided, cluster-aware way with pre-checks and progress tracking.