For most of its life, Azure Local (and Azure Stack HCI before it) meant one thing: hyperconverged. Compute and storage lived in the same box, and Storage Spaces Direct (S2D) glued it together. That model works, but if you need more capacity you buy more nodes, whether you need the CPU and memory or not.
That has changed. Azure Local now supports disaggregated deployments, where the cluster nodes run compute only and storage comes from an external SAN over Fibre Channel or iSCSI. On the Dell side, Dell Private Cloud can now deploy Azure Local on PowerEdge compute with PowerStore storage, orchestrated by the Dell Automation Platform. I spend most of my days in Azure Local validation on Dell PowerEdge hardware, so in this post I want to walk through how the pieces fit, what you need before you start, and how I validate the results with PowerShell.
If you’ve been around Azure Local (formerly Azure Stack HCI) for any length of time, you know updates are where a lot of the real-world pain lives. The October 2026 release of Azure Local (2610, solution version 12.2610.1004.50) and the changes landing in November move the update story forward in some meaningful ways, so I wanted to pull them together in one place along with how I’d adjust an update runbook.
Registering Azure Local (formerly known as Azure Stack HCI) with Azure has come leaps and bounds from the early days that I remember all too well. The process has now been simplified to a one-liner that will prompt you for Azure authentication (via Device Code) and will then start registering the node with Azure Arc to your desired resource group on your Azure subscription. Although that is great, the one thing I’m constantly asked for is guidance on how to do this via a Service Principal (app registration), or SPN for short, with the required permissions.
There are many reasons why someone would want to do this. Some of the common reasons have to deal with a company’s Azure security policy preventing them from using a device code to authenticate to Azure from an unmanaged host. Another reason may be simply because they are wanting to automate end-to-end a deployment of an Azure Local cluster (or multiple clusters).
If you find yourself in that scenario, the below will aim to provide you with some helpful code blocks that my team leverages to perform end-to-end Azure Local deployments for our validation efforts.
Recently encountered the following error while attempting to perform live migration of VMs in my cluster:
Live migration of 'Virtual Machine vm-base-A0640R33C02N01-001' failed.
Failed to get the network address for the destination node 'A0640R33C02N02': A cluster network is not available for this operation. (0x000013AB).
For some reason, valid cluster networks were excluded for migration. I was able to validate this running the following command:
To modify the virtual machine live migration exclude networks or migration network order list, you can use either Set-VMMigrationNetwork cmdlet or the Failover Cluster Manager MMC snap-in. As a best practice, the networks that should be enabled for live migration are typically the non-routable private SMB networks:
Once the live migration networks have been set accordingly, you can re-validate this on the host by running the earlier code block: