Getting Started with NC2 on AWS

The maximum number of virtual CPUs (vCPUs) that can be allocated to a single instance of a specific EC2 instance type.

A specific type of EC2 instance designed for high-performance computing, offering a large number of vCPUs and high memory capacity.

The physical location within a region where EC2 instances can be deployed. Each Availability Zone is isolated from other Availability Zones, providing fault tolerance.

The process of configuring a cluster to meet the required performance and capacity needs, taking into account the vCPU limits, instance types, and availability zones.

A storage volume that is attached to a specific EC2 instance and used to store data. EBS volumes are persistent, meaning data is not lost when the instance is stopped or terminated.

An encryption method that uses a unique key for each volume to ensure data confidentiality and integrity.

The process of running a CloudFormation template to provision and configure resources in AWS.

The minimum number of vCPUs required to create a cluster. Insufficient vCPUs may lead to cluster creation failure.

A specialized instance type designed for high-performance computing workloads that run on bare-metal hardware, providing direct access to the underlying physical resources.

A specific type of GPU designed by NVIDIA for accelerating graphics and compute-intensive tasks, often used in virtualized graphics environments.

A software driver that enables communication between the host operating system and the NVIDIA GPU to ensure proper functionality.

A set of virtualized GPUs that are allocated to virtual machines, allowing each VM to access a dedicated portion of the GPU's resources.

A mechanism in the Nutanix platform that automates the update process for the host operating system and other software components, ensuring compatibility and security.

A way to expand the size of a cluster by adding new nodes, providing additional computing resources.

A tool used to manage and access virtual machines within a cluster, offering functionality like power management and remote access.

A method of securing data by encrypting it with a unique key, ensuring confidentiality and integrity.

An AWS service providing access to instance metadata, offering both request/response (IMDSv1) and session-oriented (IMDSv2) methods. Enables interaction with instance information like ID and region.

A feature of NC2 on AWS that protects your clusters by enabling data replication to an AWS VPC.

A Nutanix account that provides access to the NC2 console, required for creating NC2 subscriptions and managing clusters.

A method of connecting on-premises datacenters with AWS VPCs securely, offering high bandwidth and low latency.

A virtual private cloud on AWS, providing a secure and isolated environment for your resources.

The process of configuring a connection between your on-premises datacenter and an AWS VPC using protocols like VPN or Direct Connect.

A specific instance type offered by AWS designed for high-performance computing, providing significant vCPU and memory capacity.

A CLI used to manage Nutanix clusters, where certain commands are disabled in an AWS environment. Examples include network management commands like adding or deleting switches, and managing the hypervisor.

A CLI designed for managing Nutanix clusters, but certain commands are not available on an AWS-hosted cluster.

A 30-day trial period for using Nutanix Cloud on all supported clouds.

The interface used to manage Nutanix Cloud (NC2) after creating a My Nutanix account.

API calls that are disabled or modified in a Nutanix cluster running on AWS. Includes actions like entering/exiting maintenance mode, fetching cluster rack configurations, and managing disks.

NC2 clusters on AWS can only have a maximum of 28 nodes in a single cluster.

NC2 does not recommend using single-node clusters for production use cases because they are not as resilient when a failure happens.

NC2 clusters do not allow mixing instances with different tenancy types (dedicated host and default tenancy). This ensures consistent resource allocation and security.

NC2 does not support the creation of heterogeneous clusters with g4dn.metal instances. This means that you can't mix g4dn.metal instances with other instance types in the same cluster.

NC2 clusters cannot share AWS subnets with other clusters. This ensures proper network isolation and security.

NC2 clusters only support IPv4 addresses and do not currently support IPv6 addresses.

Public CIDR addresses are not supported for subnets in NC2 on AWS deployments when associated with Cloud subnets. Only private CIDR addresses should be used.

Unmanaged networks are not supported in the current release of NC2. This means that network connectivity and management must be handled by Nutanix.

EBS volumes can only be attached to new clusters created with AOS 6.8 and later. Adding EBS volumes to existing clusters is not supported.

EBS volume capacity is limited to 80% of the total capacity of the instance.

With EBS volumes, only the NC2 model supports automated node replacement with similar EBS configurations. Manual replacement of EBS volumes is not supported for NC2. You can only increase EBS storage per host and cannot decrease it.

Features like Hibernate/Resume, Cluster Protect, and Software data-at-rest encryption are all supported when using EBS volumes. EBS storage is not supported in heterogeneous clusters.

When adding EBS storage to bare-metal instances during cluster creation, you can only add hosts of the same bare-metal instance type with EBS storage. EBS storage per host can be increased but cannot be decreased after deployment.

EBS volume support is available only with NCI Ultimate and AOS Ultimate licenses. You need to update the CloudFormation stack to ensure the IAM roles have the necessary permissions for EBS volume support.

EBS volumes are displayed as CLOUD-SSD tier on the Prism Element Web Console UI while they are listed as SSD-PCIe on the Prism Central UI.

AWS charges you for EBS volumes and S3 storage when the cluster is in a hibernated state. You will also be charged for any new node added for data or VM evacuation when a node becomes unhealthy.