07 — Hyper-V Failover Cluster

Overview

Property	Value
Cluster name	`hvlab-clus01`
Nodes	`hvnode01`, `hvnode02`, `hvnode03`, `hvnode04`
Storage	3 × 500 GB iSCSI CSVs from `hviscsi01`
Quorum	Cloud Witness (Azure Blob — `sthvlabwitness01`)
Domain	`azrl.mgmt`
AD OU	`OU=ClusterNodes,OU=HVLab,OU=MMS2026,DC=azrl,DC=mgmt`

Prerequisites

Before running cluster setup:

[ ] All 4 nodes domain-joined to azrl.mgmt
[ ] CNO (hvlab-clus01) pre-staged in AD (see docs/05-active-directory.md)
[ ] iSCSI disks visible on all 4 nodes with MPIO configured (see docs/06-iscsi-storage.md)
[ ] Failover-Clustering feature installed on all nodes
[ ] KCD configured for live migration (see docs/05-active-directory.md)
[ ] Cloud Witness storage account sthvlabwitness01 created with access key in Key Vault

Step 1 — Install Failover Clustering Role on All Nodes

powershell

# Run on ALL four nodes (can be parallelized with Invoke-Command)
$nodes = @("hvnode01","hvnode02","hvnode03","hvnode04")

Invoke-Command -ComputerName $nodes -ScriptBlock {
    Install-WindowsFeature -Name Failover-Clustering -IncludeManagementTools
    Install-WindowsFeature -Name Hyper-V -IncludeManagementTools
    Install-WindowsFeature -Name Hyper-V-PowerShell
}

# Reboot all nodes
Invoke-Command -ComputerName $nodes -ScriptBlock { Restart-Computer -Force }

# Wait for nodes to come back
Start-Sleep -Seconds 120
foreach ($node in $nodes) {
    while (-not (Test-Connection $node -Count 1 -Quiet)) { Start-Sleep 10 }
    Write-Host "$node is back online"
}

Step 2 — Cluster Validation

Run the cluster validation report before creating the cluster. This is required for a supported configuration and will catch any networking or storage issues.

powershell

# Run validation from hvnode01 or a management machine
Test-Cluster `
  -Node "hvnode01","hvnode02","hvnode03","hvnode04" `
  -Include "Inventory","Network","Storage","System Configuration","Hyper-V Configuration" `
  -ReportName "C:\ClusterValidation\hvlab-validation"

# Review the report
Start-Process "C:\ClusterValidation\hvlab-validation.htm"

Expected Validation Results

Test Category	Expected Result
Network	Pass — multiple networks configured
Storage	Pass — iSCSI disks visible on all nodes
Hyper-V Configuration	Pass — nested virtualization enabled
System Configuration	Warnings are acceptable (eval OS, test environment)

Do not proceed if Storage tests fail. Storage failures mean the cluster will not function correctly. Resolve iSCSI connectivity issues first.

Step 3 — Create the Cluster

powershell

# Run from hvnode01 — the CNO must already be pre-staged in AD
New-Cluster `
  -Name "hvlab-clus01" `
  -Node "hvnode01","hvnode02","hvnode03","hvnode04" `
  -StaticAddress "172.16.10.50" `
  -AdministrativeAccessPoint "ActiveDirectoryAndDns" `
  -NoStorage  # Add storage separately below

# Verify cluster creation
Get-Cluster -Name "hvlab-clus01"
Get-ClusterNode -Cluster "hvlab-clus01"

The cluster IP 172.16.10.50 is the Cluster Name Object (CNO) IP on the Mgmt network. This is used for Failover Cluster Manager connections and WAC/SCVMM cluster management.

Step 4 — Add Storage to the Cluster

powershell

# Add the iSCSI disks to the cluster
# First, identify the disk numbers (run on hvnode01)
Get-Disk | Where-Object BusType -eq "iSCSI" |
    Select-Object Number, Size, FriendlyName | Sort-Object Size

# Add all available storage to the cluster
Get-ClusterAvailableDisk -Cluster "hvlab-clus01" |
    Add-ClusterDisk

# Verify disks are added
Get-ClusterResource -Cluster "hvlab-clus01" |
    Where-Object ResourceType -eq "Physical Disk"

Step 5 — Convert Disks to Cluster Shared Volumes

powershell

# Get the cluster disk resources (exclude the small quorum disk)
$clusterDisks = Get-ClusterResource -Cluster "hvlab-clus01" |
    Where-Object ResourceType -eq "Physical Disk"

# Convert the 3 × 500 GB disks to CSVs (skip the 10 GB quorum disk)
foreach ($disk in $clusterDisks) {
    $physDisk = Get-ClusterParameter -InputObject $disk |
                Where-Object Name -eq "DiskSize"
    if ($physDisk.Value -gt 100GB) {
        Add-ClusterSharedVolume -InputObject $disk
        Write-Host "Added CSV: $($disk.Name)"
    }
}

# Rename CSVs to meaningful names
$csvs = Get-ClusterSharedVolume -Cluster "hvlab-clus01"
$csvs | Format-Table Name, State, SharedVolumeInfo

# Rename (adjust names based on what was created)
(Get-ClusterSharedVolume -Cluster "hvlab-clus01" -Name "Cluster Disk 2").Name = "CSV-Vol1"
(Get-ClusterSharedVolume -Cluster "hvlab-clus01" -Name "Cluster Disk 3").Name = "CSV-Vol2"
(Get-ClusterSharedVolume -Cluster "hvlab-clus01" -Name "Cluster Disk 4").Name = "CSV-Vol3"

# Verify CSV mount points (accessible as C:\ClusterStorage\Volume1, etc.)
Get-ClusterSharedVolume | Select-Object Name, State,
    @{N="Path";E={$_.SharedVolumeInfo.FriendlyVolumeName}}

Step 6 — Configure Cloud Witness

powershell

# Retrieve the witness storage account key from Key Vault
$witnessKey = az keyvault secret show `
  --vault-name "kv-hvlab-mms26-eus-01" `
  --name "hvlab-witness-storage-key" `
  --query value -o tsv

# Set Cloud Witness
Set-ClusterQuorum `
  -Cluster "hvlab-clus01" `
  -CloudWitness `
  -AccountName "sthvlabwitness01" `
  -AccessKey $witnessKey `
  -Endpoint "core.windows.net"

# Verify quorum configuration
Get-ClusterQuorum -Cluster "hvlab-clus01"

See docs/08-cloud-witness.md for details on Cloud Witness behavior and graceful degradation.

Step 7 — Configure Cluster Networks

After cluster creation, set the correct roles for each network:

powershell

# View current cluster networks
Get-ClusterNetwork -Cluster "hvlab-clus01" | Select-Object Name, Address, Role

# Set network roles
# Role 0 = Do not allow cluster network communication
# Role 1 = Allow cluster network communication only
# Role 3 = Allow cluster network communication and client connectivity

# Mgmt network — allow both cluster and client
(Get-ClusterNetwork -Cluster "hvlab-clus01" |
    Where-Object Address -eq "172.16.10.0").Role = 3

# Migration network — cluster only
(Get-ClusterNetwork -Cluster "hvlab-clus01" |
    Where-Object Address -eq "172.16.20.0").Role = 1

# Heartbeat network — cluster only
(Get-ClusterNetwork -Cluster "hvlab-clus01" |
    Where-Object Address -eq "172.16.40.0").Role = 1

# Storage network — do NOT use for cluster communication
(Get-ClusterNetwork -Cluster "hvlab-clus01" |
    Where-Object Address -eq "172.16.30.0").Role = 0

# Rename networks for clarity
(Get-ClusterNetwork -Cluster "hvlab-clus01" | Where-Object Address -eq "172.16.10.0").Name = "Cluster-Mgmt"
(Get-ClusterNetwork -Cluster "hvlab-clus01" | Where-Object Address -eq "172.16.20.0").Name = "Cluster-Migration"
(Get-ClusterNetwork -Cluster "hvlab-clus01" | Where-Object Address -eq "172.16.30.0").Name = "Cluster-Storage"
(Get-ClusterNetwork -Cluster "hvlab-clus01" | Where-Object Address -eq "172.16.40.0").Name = "Cluster-Heartbeat"

Step 8 — Configure Live Migration

powershell

# On all cluster nodes — configure live migration settings
Invoke-Command -ComputerName @("hvnode01","hvnode02","hvnode03","hvnode04") -ScriptBlock {
    # Enable live migration
    Enable-VMMigration

    # Use Kerberos authentication (KCD configured in AD step)
    Set-VMHost -VirtualMachineMigrationAuthenticationType Kerberos

    # Limit concurrent live migrations to 4 (adequate for demo, avoids saturation)
    Set-VMHost -MaximumVirtualMachineMigrations 4

    # Set migration network to the Migration vSwitch
    Add-VMMigrationNetwork -Subnet "172.16.20.0/24" -Priority 1

    # Ensure the Mgmt network is lower priority for migration
    Add-VMMigrationNetwork -Subnet "172.16.10.0/24" -Priority 5
}

Step 9 — VM Placement Rules for Demo

Configure preferred owners and failover policies for demo VMs to ensure predictable behavior during the presentation:

powershell

# All demo cluster roles should have all nodes as possible owners
# Set preferred owners for demo scenarios

# Example: Create a demo VM resource with controlled placement
# This is run after demo VMs are added to the cluster

$demoVMs = Get-ClusterGroup -Cluster "hvlab-clus01" |
    Where-Object { $_.GroupType -eq "VirtualMachine" }

foreach ($vm in $demoVMs) {
    # Set all nodes as preferred owners (equal priority)
    Set-ClusterOwnerNode -InputObject $vm `
        -Owners "hvnode01","hvnode02","hvnode03","hvnode04"
    Write-Host "Set owners for: $($vm.Name)"
}

# Configure anti-affinity for better demo — spread VMs across nodes
# (SCVMM handles this better — use SCVMM placement for demo VMs)

Verification

After completing all steps, run a final cluster health check:

powershell

# Overall cluster health
Get-Cluster -Name "hvlab-clus01" | Select-Object Name, QuorumType, SharedVolumesRoot

# Node health
Get-ClusterNode -Cluster "hvlab-clus01" | Select-Object Name, State, ID

# Network health
Get-ClusterNetwork -Cluster "hvlab-clus01" | Select-Object Name, Role, State

# CSV health
Get-ClusterSharedVolume -Cluster "hvlab-clus01" | Select-Object Name, State,
    @{N="Owner";E={$_.OwnerNode.Name}},
    @{N="Path";E={$_.SharedVolumeInfo.FriendlyVolumeName}}

# Quorum health
Get-ClusterQuorum -Cluster "hvlab-clus01"

# Live migration test (migrate hvnode01's VMs to hvnode02)
# Get-ClusterGroup -Cluster "hvlab-clus01" | Where-Object OwnerNode -eq "hvnode01" |
#     Move-ClusterVirtualMachineRole -Node "hvnode02"

Expected Healthy State

Cluster:  hvlab-clus01  QuorumType: CloudWitness
Nodes:    hvnode01-04   State: Up
Networks: Cluster-Mgmt (Role: 3), Cluster-Migration (Role: 1),
          Cluster-Heartbeat (Role: 1), Cluster-Storage (Role: 0)
CSVs:     CSV-Vol1, CSV-Vol2, CSV-Vol3  State: Online
Quorum:   CloudWitness  AccountName: sthvlabwitness01

Troubleshooting

Cluster Validation Fails on Storage

powershell

# Confirm all nodes see the same disks
Invoke-Command -ComputerName @("hvnode01","hvnode02","hvnode03","hvnode04") -ScriptBlock {
    Get-Disk | Where-Object BusType -eq "iSCSI" |
        Select-Object Number, SerialNumber, Size | Sort-Object SerialNumber
}
# Serial numbers should match across all nodes for each disk

Node Shows as Down

powershell

# From a working node
Test-Connection hvnode02 -Count 3

# Check cluster node status and event log
Get-ClusterNode "hvnode02" | Select-Object Name, State, StatusInformation
Get-WinEvent -ComputerName hvnode02 -LogName "Microsoft-Windows-FailoverClustering/Operational" `
    -MaxEvents 20 | Select-Object TimeCreated, Message

CSV Not Coming Online

powershell

# Check which node owns the CSV and the underlying disk
Get-ClusterSharedVolume -Cluster "hvlab-clus01" | Select-Object Name, State, OwnerNode

# Check the physical disk resource
Get-ClusterResource -Cluster "hvlab-clus01" |
    Where-Object ResourceType -eq "Physical Disk" |
    Select-Object Name, State, OwnerGroup

07 — Hyper-V Failover Cluster ​

Overview ​

Prerequisites ​

Step 1 — Install Failover Clustering Role on All Nodes ​

Step 2 — Cluster Validation ​

Expected Validation Results ​

Step 3 — Create the Cluster ​

Step 4 — Add Storage to the Cluster ​

Step 5 — Convert Disks to Cluster Shared Volumes ​

Step 6 — Configure Cloud Witness ​

Step 7 — Configure Cluster Networks ​

Step 8 — Configure Live Migration ​

Step 9 — VM Placement Rules for Demo ​

Verification ​

Expected Healthy State ​

Troubleshooting ​

Cluster Validation Fails on Storage ​

Node Shows as Down ​

CSV Not Coming Online ​

07 — Hyper-V Failover Cluster

Overview

Prerequisites

Step 1 — Install Failover Clustering Role on All Nodes

Step 2 — Cluster Validation

Expected Validation Results

Step 3 — Create the Cluster

Step 4 — Add Storage to the Cluster

Step 5 — Convert Disks to Cluster Shared Volumes

Step 6 — Configure Cloud Witness

Step 7 — Configure Cluster Networks

Step 8 — Configure Live Migration

Step 9 — VM Placement Rules for Demo

Verification

Expected Healthy State

Troubleshooting

Cluster Validation Fails on Storage

Node Shows as Down

CSV Not Coming Online