This article provides architectural considerations, deployment options, and scaling guidance to migrate virtual machines from VMware ESXi to a Nutanix AHV cluster connected to a Everpure FlashArray using Nutanix Move.
Overview
Nutanix Move is a migration tool designed to simplify the movement of virtual machines from supported hypervisors to the Nutanix AHV hypervisor. It enables organizations to migrate workloads with minimal downtime by using staged data synchronization followed by a controlled cutover process.
This article focuses specifically on migrations from VMware ESXi to Nutanix AHV, as this is the most common migration scenario. The assumption is that the Nutanix AHV cluster is connected to a Everpure FlashArray as external storage.
Nutanix Move also supports migrations from other hypervisors and environments. For details on supported source platforms and additional migration workflows, refer to the official Nutanix Move documentation.
For step-by-step migration procedures, refer to the Nutanix support site.
A sample migration video is shown at the end of this article.
Pre-requisites
-
Minimum Nutanix Move version: 6.1.0
-
Minimum FlashArray Purity//FA version: 6.10.3
-
Minimum Nutanix version: Nutanix AOS 7.5.0 / Nutanix AHV 11
-
Move network connectivity: between source VMware vCenter environment and target Nutanix NCP cluster.
Nutanix Move Architecture and Deployment Models
Nutanix Move is deployed as a virtual appliance and can be installed in either the source or destination environment. However, it is recommended by Nutanix that Move is installed on the target cluster.
Nutanix Move can be downloaded from the Prism Central Marketplace or from the Nutanix Support portal as an OVA or disk image.
Supported Deployment Locations:
-
Deployed on VMware ESXi (vCenter-managed): The Move appliance runs as a VM on ESXi and communicates with vCenter, source ESXi hosts, and the destination AHV cluster.
-
Deployed on Nutanix AHV: The Move appliance runs as an AHV VM and connects to vCenter and the ESXi source environment remotely. This is the recommended option for most customer migrations.
Both deployment models are fully supported for ESXi-to-AHV migrations. Deployment location does not affect migration capabilities, but external influences such as network routing and throughput may vary depending on environment design.
Network and Workload Considerations
-
Guest VM connectivity to the Move appliance is required. Nutanix Move relies on in-guest communication during migration. All VMs being migrated must be able to establish network connectivity to the Move appliance.
-
VMs deployed in De-militarized Zones (DMZ), isolated networks, or environments with restrictive firewall policies may require temporary routing or firewall changes. If connectivity cannot be established, those VMs cannot be migrated using Move.
-
Not all workloads are eligible for migration. Certain workloads like vendor-supplied virtual appliances, infrastructure VMs, or appliances with restrictive licensing or unsupported guest configurations — are not supported for migration with Nutanix Move and must be re-deployed on AHV rather than migrated.
For detailed networking requirements, including supported operating systems and connectivity expectations, refer to the official Nutanix Move documentation.
Storage Considerations for Everpure FlashArray
When Everpure FlashArray is used in the source and/or destination environment, any FlashArray-backed datastore or volume is a suitable location for deploying the Nutanix Move appliance.
No special FlashArray volume configuration is required beyond standard best practices for the hypervisor in use. However, it is important to ensure that FlashArray system limits and feature capabilities are considered as part of migration planning.
In particular, object counts (such as volumes, snapshots, hosts, and connections) can increase during migration activities and should be monitored to avoid approaching platform limits. Tracking object counts is especially important in large-scale or highly parallel migration scenarios.
For more detailed information on maximum object counts, supported features, and platform capabilities, refer to the official Everpure FlashArray documentation covering system limits and feature support. This information can help validate that the FlashArray is appropriately sized and configured to support the migration workflow.
Nutanix Move does not leverage FlashArray replication or array-based data movement to perform migrations. Instead, VM data is read from the source volumes and transferred over the front-end storage and network paths as part of the Move migration workflow. As a result, migration activity generates standard host I/O traffic on the array and should be considered alongside other workloads when planning migration timing and scale.
When Everpure FlashArray is used as the destination storage platform for Nutanix AHV, data reduction (deduplication and compression) is applied opportunistically as VMs are migrated from ESXi to Nutanix AHV, based on overall FlashArray processing workload.
Customers should generally expect to see similar data reduction ratios on Nutanix AHV as those previously observed on VMware ESXi for migrated VMs. If the array is under sustained heavy write operations, customers should ensure that sufficient swing space is available during migration activities.
In environments with low overall utilization, little additional swing space is typically required. However, when the array is running significant I/O from other workloads, data reduction may not always be applied inline in real time. In these cases, data may be written temporarily without full reduction and will be deduplicated and compressed later once FlashArray processing workload decreases.
When Nutanix AHV is used as the migration destination, selecting Everpure FlashArray as the target storage is done by choosing the appropriate FlashArray-backed storage container within Move. No additional configuration or integration steps are required within Nutanix Move beyond selecting the desired container.
If the FlashArray has not yet been configured as external storage for the Nutanix AHV cluster, it must first be set up by following the Nutanix and Everpure FlashArray Quick Start Guide before it can be selected as a migration target.
Migration Workflow and Parallelism
Nutanix Move supports the migration of multiple virtual machines in parallel within a single migration plan. The number of VMs that can be migrated concurrently depends on several factors, including VM size, disk activity, number of disks, network bandwidth, and the resources allocated to the Move appliance.
To copy VM data, Nutanix Move takes VMware snapshots on the source ESXi environment. An initial snapshot is taken to seed the bulk of the VM data, followed by one or more incremental snapshots to capture changed blocks leading up to the final cutover. The frequency of incremental snapshots can be set in the Move appliance, with the default value being every 10 minutes.
For VMs running on FlashArray-backed VMware Virtual Volumes (vVols), these snapshots are implemented as array-based snapshots on the storage platform, which provides better performance and lower impact to running workloads.
In contrast, VMFS-based snapshots rely on hypervisor-managed delta disks and can introduce additional overhead, particularly for larger VMs with high data change rates. As a result, workloads with significant write activity may experience increased snapshot-related impact when multiple VMFS-based snapshots are taken during the migration process.
General Guidance:
-
A single Move instance supports a minimum of 1 to a maximum of 100 VMs per migration plan, depending on workload characteristics.
-
For large environments (200+ VMs), it is recommended to use multiple migration plans and/or multiple Move instances to improve predictability and reduce cutover risk.
-
Migration plans should be treated as logical “waves” of workloads rather than a single bulk operation.
-
When migrating VMs backed by VMFS datastores, pay attention to write-intensive workloads, as multiple ESXi-based snapshots may be taken during the migration process. Where available, VMware Virtual Volumes (vVols) can reduce snapshot overhead by leveraging FlashArray-based snapshots.
Using Multiple Move Instances:
For higher aggregate throughput or large-scale migrations (several hundred and more VMs), multiple Nutanix Move appliances can be deployed in parallel, especially if the migration spans multiple storage arrays and/or vCenters. Each Move instance operates independently and manages its own migration plans. One recommended approach is to deploy a Move appliance per ESXi host as this avoids potential migration overlaps.
Using multiple Move instances provides:
-
Increased overall migration throughput
-
Improved fault isolation between migration waves
-
Greater flexibility in scheduling and cutover timing
Move Appliance Resource Sizing:
-
The Nutanix Move appliance is deployed with default virtual CPU and memory allocations that are sufficient for most small to medium migration scenarios. In some cases, adding additional vCPU and memory may provide incremental performance improvements.
-
Resource scaling should be validated through test migrations. Performance improvements are not linear, and migration throughput is often constrained by network bandwidth or source and destination storage performance rather than by Move appliance resources.
-
When the source or destination storage platform is also running production workloads, the additional read and write activity generated by Move can impact those workloads. In high storage load scenarios, data seeding and migrations should be scheduled during off-hours.
-
To increase vCPU or memory, the Move appliance must be powered off and the additional resources added before restarting the appliance.
Nutanix Move Migration Demo Video