At Microsoft Ignite 2025 we announced our first public preview of Windows Admin Center: Virtualization Mode. The reception has been nothing short of overwhelming! Many of you have reached out to test, ask questions, and provide feedback.

Now, only a few months later, we’ve made some substantial improvements, and it’s time for our next round of feedback before we head towards general availability. This article will take you through the various improvements in Public Preview 2 so you can test this out in your environment and provide some additional feedback.

Installation and Upgrades

From the start, we set out to avoid long, prerequisite-heavy installs. In Public Preview 1, vMode installs in under five minutes and has just one prerequisite (the Visual C++ Redistributable). Our Installing vMode blog shows a full install in 2:12.

Of course, we understand that installation is a one-time task and vMode will need periodic updates—so we’ve remained focused on making it easy to run and maintain. Here are the installation and upgrade improvements you’ll see in Public Preview 2.

Unattended Installation

INI files may feel a little old‑school, but in the world of unattended Windows installations they’re quietly powerful. They give you a lightweight, deterministic way to drive setup without the overhead of more complex provisioning systems. In PP2, we support vMode installation through INI files. INI files are dead‑simple: sections, keys, values. That simplicity means:

• No parsing ambiguity
• No schema validation overhead
• No tool dependencies
• Easy human readability

To do this at an elevated command line, first, get an encrypted version of the PostGreSQL password you'll want to use

"YourPassword" | ConvertTo-SecureString -AsPlainText -Force | ConvertFrom-SecureString

Next, create your INI file. You only need to specify the functionality you want to override (all other keys are optional by default). Note: The INI file must have the [AppSettings] section and the PostgreSQL properties. For the password, use the output of the ConvertTo-SecureString cmdlet we just ran.

[AppSettings]
PostgreSQLUsername=postgres
PostgreSQLPassword=01000000d08c9ddf0115d1118c7a00c04fc297eb...
PostgreSQLPort=5432

Now, pass in the configuration file with the /ConfigFile parameter at a command prompt

WindowsAdminCenterSetup.exe /VERYSILENT /ConfigFile="C:\deploy\wac-config.ini"

Here's a video showing a simple unattended installation:

Upgrades and Resource Health

To upgrade from Public Preview 1 to Public Preview 2, just run the PP2 installer on your vMode system. The experience is the same as a fresh install (but it keeps the previous settings).

During the upgrade process, the database is upgraded automatically—no scripts, custom steps, or guesswork about upgrade order. Just double-click the installer and go!

After you sign back into vMode, the Deployment Status page will show upgrade activity. If the release requires newer machine agents, vMode upgrades them automatically. When that completes, the upgrade is done.

And if anything goes wrong (or more likely for peace of mind 😊), use the Resource Health page to monitor agent status, expiring certificates, and more. Open it by selecting the Virtualization Mode machine in the navigation hierarchy. This page also lets you review and download logs to share with support.

Take a look for yourself at how easy the upgrade process is:

Add Resource Wizard

We designed vMode to feel intuitive; this means one onboarding wizard that covers the full set of onboarding scenarios rather than separate wizards for each workflow.

Public Preview 2 includes many improvements to the Add Resource wizard, outlined below.

All Profiles: Better detection of existing configuration

Existing Agents

We've improved our intelligence when you add nodes to the Add Resource wizard. If the system being added already has an agent and we think it could be hosted by another vMode installation, we notify you and block progression through the wizard until resolved. Whether due to misconfigured DNS or a typo when you entered the name, we don't want to accidentally "commandeer" an already managed system from another vMode instance.

Adding a host with an existing agent

Detecting Cluster Membership

When you add a cluster member to the Add Resource wizard, we'll automatically detect and add the other nodes in that cluster. This is a simple improvement designed to prevent mistakes.

Compute profile: Existing Hyper-V Clusters

A key challenge in IT infrastructure is keeping systems in their expected state. Throughout the wizard, we verify that all nodes in a cluster are consistently configured and healthy—including updates and pending reboots.

In Public Preview 2, we now install available updates and resolve pending reboots (whether because of feature installation or updates). If needed, we perform a rolling reboot of cluster nodes and drain VMs from each node first to avoid disruption.

Note: In the future, we'll assess cluster capacity using Cluster Aware Updating prior to beginning the operation.

However, in PP2, if the VMs do not evacuate the host for any reason (for example, insufficient cluster capacity), the wizard will not reboot the host and will generate a warning. This can lead to longer onboarding times as we do not proactively check for capacity.

Compute profile: Create Hyper-V Clusters

In Public Preview 1, we supported onboarding existing Hyper-V clusters. In Public Preview 2, the same Add Resource wizard can create a new compute cluster and setup all the nodes. To do this, just add standalone nodes and select the clustered option.

Create Hyper-V cluster from standalone nodes

On the clustering page, you'll need to enter a name for the cluster and use DHCP or a static IP address for it.

Compute or Storage profiles: Networking configuration page

The networking step is where you define Network ATC intent templates for Management, Compute, and Storage traffic on your NICs.

Intent Templates now preserved

In Public Preview 1, the Add Resource wizard did not save intent templates, so you had to recreate them each time you onboarded new nodes or clusters. In Public Preview 2, saved templates are preserved and ready to reuse directly from the wizard.

All NICs now shown

We also fixed a common issue in adapter discovery. Previously, only connected adapters appeared, which hid disconnected but physically present NICs. In Public Preview 2, all physical Ethernet adapters appear regardless of connection state, so you can see your full hardware inventory before configuring intents.

Previously, only connected adapters would be shown in the wizard which meant that disconnected, but physically present NICs would be hidden. This really cleaned up the UI if you had a bunch of NICs you don't plan to use, but could be quite confusing if you were expecting to see a NIC in the list and didn't!

Default values for overrides prepopulated

When you configure an intent, default Network ATC override values are now prepopulated so you can see exactly what will be applied. If a default works for your environment, leave it unchanged. Defaults are not submitted as overrides, which keeps the configuration clean. Only values you change are sent as overrides.

Automatic override configuration for test environments

Many of you likely test your configuration in VMs before deployment. In those VM testing scenarios, Network ATC requires the NetworkDirect (RDMA) adapter property override to be disabled so you can validate your intents.

To simplify this, vMode automatically sets this override for you when virtual machines are detected. You do not need to change anything manually.

Onboarding validation improved

The Add Resource workflow now validates your network intent successfully applied before onboarding completes. You can see this in the Deployment Status tab once the workflow begins. Now, if your networking cannot be setup properly, you'll be notified upfront and the workflow will fail so you can address the issue.

Compute profile: Storage configuration page

This screen allows a user to onboard or configure storage for the Hyper-V virtual machines. In our first release, we had limited support for SANs and NAS devices. Now in PP2, you can provision a brand new LUN to the nodes in your cluster and the wizard will turn it into a CSV. When you arrive at this page, LUNs will be sorted into three different categories:

• Ready – This LUN is already a CSV in the cluster (applies to existing clusters only)
• Auto configure – If you specify a name for the LUN, we will set it up for you
• Manual – The LUN requires manual intervention to meet the requirements of a failover cluster

If you select a LUN from the auto configure section, vMode will:

1. Initialize and online the disk
2. Create a GPT partition and format it with NTFS
3. Add the disk to the failover cluster
4. Add the disk as a Cluster Shared Volume with the name you specify
5. Ensure the file system matches the CSV name (e.g. CSV01 C:\ClusterStorage\CSV01)
6. CSV Block Cache is configured to the recommended value (you can always increase this after deployment)

Here's a video showing SAN storage configuration while onboarding a Hyper-V cluster to vMode. In the video, you'll find that there are three LUNs provisioned. The first LUN in the 'Ready' state is already setup as a CSV in the cluster; the second LUN in the 'auto-configure' state is a raw disk with no configuration on it; the third LUN is in the 'Manual' state because only 3 of the 4 nodes in the cluster can see it. Once that issue is resolved, LUNs 2 and 3 are both in the 'Auto-configure' state and provisioned as CSVs by the wizard.

Compute profile: Compute configuration page

This page configures Hyper-V host settings but the options were greyed out in Public Preview 1 with the intent of showing you what would be there in the future – The future has arrived!

Here’s how we handle each of the options on this page:

• Remove non-essential features
  If selected, we’ll remove any unused Windows Server features. This reduces patching time and reboots by eliminating unnecessary OS components. For details, see our previous compute configuration blog.
  
• Enable Enhanced Session Mode
  Enhanced Session Mode provides a rich, desktop-like experience, letting you dynamically resize the VM window, copy and paste between the host and VM, map host hardware into the VM, and play or record audio through the host.
  
  In Hyper-V, this is two separate configurations on each host. In vMode, we make this one simple box to select enabling both the user and computer configuration.
• Configure concurrent migrations
  This is another “two-for-one.” Configuring migrations is straightforward whether it be a live migration or storage migration—just choose 1, 2, or 4 in the wizard, and it applies to all machines. You can later adjust this setting in WAC Hyper-V as needed.

• Default virtual machine path
  Another “two-for-one.” In Hyper-V, you set separate paths for Virtual Machines and Virtual Hard Disks. With vMode, a single setting configures both to the same location. If you configured a storage path earlier, you can use that here as well.

Note: Enhanced Session Mode (user and server settings), Configure concurrent migrations (storage and live migration), and default virtual machine paths (VHD and VM), can each be verified in the Hyper-V settings for each server specified in the wizard

Storage profile: Create a Scale-out File Server

In public preview 2, if you select the storage profile and add a cluster to the wizard, you can now create a Scale-out File Server.  A Scale-out file server  (SOFS) provides continuously available file shares for Hyper-V virtual machines. There are many benefits to using a SOFS share and we encourage you to check the link above for more information.

In this video, we use an existing cluster with a CSV to create and manage a SOFS target. We also create the share for you making this a very simple experience. Over time, we'll be adding even more functionality to help you create the SOFS cluster, CSV configuration, etc.

Views

Storage View improvements

Storage is one of our three "views" of the navigation that allows you to focus on a specific topic and see the objects that are important for the task you're thinking about at that time. In the Storage view, we have several different types of storage objects that can be added and each have their own icons. For example, you might have Windows-based scale-out file servers (shown in the video above) or your own storage vendors devices. Over time, we'll help connect the storage with the VMs and resources they're providing so it's easy to manage the storage. We're also working with Storage vendors to enhance the management capabilities through vMode.

If you're interested in managing your storage through vMode, please speak with your Storage vendor.

Network View improvements

Now, Network ATC will be a visible, manageable part of the experience rather than something you configure through PowerShell and then hope is right. You’ll be able to manage your network intent templates through the Network View by viewing, creating, editing, and deleting your network intent templates all through a single pane of glass.

Global Search

vMode supports up to 1,000 hosts and 25,000 virtual machines. With all those systems to manage, it can be quite cumbersome to find the proverbial “VM in the vMode haystack.” Even in smaller environments, it’s unreasonable to remember where every virtual machine, template, host, storage device, etc. might be living in the navigation.

To make searching for managed objects easier, we’ve improved the Global Search capability to include Profile, Object, and Hierarchical search.

Profile Search

Objects in vMode have a profile, for example, Compute, Storage, or Networking. You can narrow down that match your search criteria by selecting the profile buttons in the search bar. If you want all compute, storage, or networking devices returned, just ensure that the "All" button is selected. As a reminder, the profile directly correlates to the view where you find the objects in the navigation.

Object Search

This capability allows you to search for a specific type of object. Imagine you’re looking for a specific virtual machine but the results kept showing you other objects (like clusters and hosts) that match the search criteria.

In this case, you can tell Global Search to ignore anything but the specified object types. Object types are specified in between square brackets. Next, you just add your search criteria and BAM! Your search is whittled down to only the appropriate object types.

We recognize that our specific object names might be difficult to remember and type, so we display the object names to the right of the search criteria or when you type an opening square bracket on-screen to make selecting the object name more easily.

Hierarchical Search

This type of search enables you to search in a specific part of the navigation hierarchy. For example, you might know that the system you’re looking for is in a specific resource group or cluster. You wouldn’t want the results being muddied up by objects that appear in locations in the navigation.

In this case, you use the forward slash (/) delimiter to indicate a hierarchical search. You can put the full path as seen in navigation with the forward slash separating each new object you want to search under. Next, just put a colon and your search criteria.

Virtual Machine Management

VM Templates

Hyper-V has never had native VM Template support but with vMode, that’s changing. Coming in PP2, you’ll have right-click options that let you convert an existing VM to a template. Once converted to a template, you can either convert that template back to a VM if you want to make any modifications, or you can deploy the template as a new VM.

Behind the scenes

When you convert a VM to a template, we modify some of the VM metadata to ensure that it’s not tampered with and is easy to identify. For example, the VM will

• have a special icon to make it clear that it’s a template
• be renamed to have a prefix of {Template}{Read-only}

Note: This doesn’t change the name of the VM inside the guest.

Not…behind the scenes

We don’t ask you to place this VM in any special place like a separate library machine. We have two primary reasons that we DON’T ask you to do this. First, copying the template to or from the library adds significant time to the process. Second, having a dedicated machine to act as a library means you have one more machine you have to deploy, maintain, update, secure, etc. Your day is busy enough!

Power management in object hierarchy

As part of the feedback we received in PP1, we’ve updated the VMs in the navigation hierarchy to have power state indicators on their icons. This is simply a convenience to quickly identify whether a VM is on, off, starting, etc. Of course, the right-click and menu options update automatically whenever the power states change as well.

Hyper-V Replica

Hyper-V Replica provides built-in disaster recovery for virtual machines and removes the high cost of implementing DR. It’s been around since Windows Server 2012 but has not received a major management upgrade for a long time. This all changes with vMode!

Hyper-V Replica: Cluster support

Our initial release in PP1 only supported standalone host integration. Now, we have full support for the Hyper-V Replica cluster broker. The broker lets you target another Hyper-V cluster for failover and gives your entire Hyper‑V cluster a single, consistent identity for receiving replication traffic. It ensures that virtual machine replication keeps working smoothly even when VMs move between cluster nodes. In PP2, we automatically detect if you need the cluster broker and and guides you accordingly.

Hyper-V Replica: Post-failover IP injection

This is one of the hidden gems that makes DR failover work across customer environments. When failing over from one datacenter to another, the network subnets often change. If you don’t use DHCP in your environment for your server machines, you need a way to automatically update the IP address when it gets to the failover site.

Now you can pre-configure the failover network settings for each VM's network adapter, specifying exactly what IP configuration the VM should receive when it arrives at the DR site. This means the IP is injected into the guest OS automatically during failover.

Hyper-V Replica: Extended replication

What’s worse than one data center failure? Two data center failures! Often, we hear that failover sites are near the original site. This has the advantage of making failover simpler and significantly faster. However, if the disaster is a regional issue (for example a power grid outage), then you might want to extend that replication to a tertiary site.

In PP2, you can configure your replication settings to extend from your failover site to a 3^rd site. The settings interface allows you to configure and manage these multi-hop replication chains. Of course this could be another local data center, or up to Azure VM..

Hyper-V Replica: Replication and failover improvements

Reverse Replication

Failover gets you through the emergency. Reverse replication gets you back home. For example, before failover, Site A was sending data to Site B. After failover, your VM is running at Site B. Reverse replication simply flips the arrow, now Site B sends data back to Site A. Once everything is caught up, you can do a planned failover back to Site A and you're back to where you started.

This functionality is now available in PP2.

Test Failover

A disaster recovery plan you've never tested is just a hope. Test failover lets you validate that your replicated VM actually works at the DR site, without disrupting production replication or affecting your live workloads.

When you run a test failover, Hyper-V creates a temporary copy of the replica VM from a recovery point you choose. This test VM runs completely isolated — your production VM keeps replicating normally, and the test VM doesn't interfere with it. You can log in and confirm your DR plan works as expected.

Now you can initiate a test failover from the replication action bar on the replica side. You just pick a recovery point - latest, standard, or application-consistent and HVR spins up the test VM. When you're done validating, clean up by stopping the test failover, which will automatically remove the temporary VM.

Summary

If you're migrating to a new virtualization platform, you want a management tool that makes things simple whether that be installation and upgrade or disaster recovery. We've made some great strides in Public Preview 2 including new Add Resource scenarios, VM Templates, DR Failover, vMode upgrades and much more though we're not done yet! Give Public Preview 2 some testing and give us your feedback on our GitHub page. We want to hear from you so we can make the highest quality virtualization platform management tool!

 

Thanks for reading,

Dan "Still Virtualizin'" Cuomo