Skip to content

Deploy-First Migration (Carbonite Migrate)

Build the destination VM on Azure Local first, then use Carbonite Migrate to replicate the source workload directly into it.

Scenario Overview

Deploy-First with Carbonite Migrate is the right choice when you need low-downtime, agent-based OS-level replication into a pre-provisioned Azure Local VM. Instead of relying on hypervisor-native replication or a two-hop staging path, you build the target VM first, install Carbonite agents on both endpoints, and let Carbonite drive continuous block-level replication until cutover.

Overview

Deploy-First is a build-first migration strategy. The destination VM exists on Azure Local before any migration activity starts. Carbonite Migrate then handles OS-level replication from the Nutanix source directly to the pre-built target — no intermediate staging server, no hypervisor API dependency.

This makes it a strong fit for:

  • Mixed or legacy estates where hypervisor-native migration is unavailable or impractical
  • Workloads that need live cutover with minimal production downtime
  • Environments where VM re-IP or OS-level state preservation is required, but the overhead of a full two-hop pipeline is not justified

Scenario Pages

  • Prerequisites — Carbonite licensing, management server, agent requirements, and network ports
  • Architecture — Deploy-First with Carbonite component model and data flow
  • Runbook — Step-by-step Carbonite Migrate execution
  • Validation & Checklist — Cutover validation, rollback guidance, and sign-off checklist

Migration process at a glance

Phase What happens
Provision Target VM built and baselined on Azure Local before migration starts
Agent install Carbonite Migrate agent installed on source Nutanix VM and target Azure Local VM
Initial mirror Carbonite performs full block-level synchronization from source to target
Continuous replication Changed-block replication keeps source and target in sync until cutover
Cutover Source writes stopped, final sync completes, execution switches to target VM
Validation Application owner validates target; source VM held for rollback window

When to Use This Approach

Use Case Why Deploy-First + Carbonite
Low-downtime OS-level migration required Continuous replication minimizes the cutover window
Hypervisor APIs unavailable or not preferred Agent-based — no Nutanix or VMware dependency
Mixed OS estate across Windows and Linux Carbonite supports both
VM right-sizing needed alongside migration Target VM is provisioned with correct sizing before migration starts
No two-hop staging infrastructure available Direct source-to-target path, no intermediate Hyper-V host required

Not Suitable For

  • Workloads requiring a full hypervisor-backed snapshot or point-in-time restore model during migration
  • Environments where agent installation on source VMs is not permitted (use Veeam or HYCU instead)
  • Pure file-server or application-layer workloads where a simpler approach is preferred (see Alternative Migration Methods)

Resources

Alternative approaches

  • If you need a two-hop replication path with built-in re-IP, see Veeam
  • If you want a simpler Nutanix-native backup/restore workflow, see HYCU
  • If you want to stay inside an existing Commvault operating model, see Commvault
  • For file-server or app-native migration without Carbonite, see Alternative Migration Methods