Skip to content

Validation

PowerShell Run on: Mgmt Workstation

Overview

After deploying and configuring the SOFS, validate that everything works before onboarding AVD session hosts. This page covers automated validation, manual verification steps, and failover testing.

Per-Tool Testing

Each deployment tool has its own unit/integration tests that can be run before deployment:

PowerShell (Pester 5)

Invoke-Pester -Path ./tests -Output Detailed

Runs 23 test suites covering config parsing, schema validation, VM parameter generation, storage path mapping, FSRM quotas, Cloud Cache, and more.

Terraform

cd src/terraform
terraform init
terraform test

Validates AVM module references, variable mapping, domain join extension, per-VM storage paths, and single/triple layout conditional logic.

Ansible (Molecule)

cd src/ansible
molecule test

Runs the full Molecule test suite — lint, syntax, converge, idempotence, verify — against a local Docker test matrix.

Bicep

az bicep build --file src/bicep/main.bicep

Validates Bicep syntax and AVM module references. Successful compilation confirms the template is structurally valid.

ARM

cd src/arm
.\build-arm.ps1

Recompiles from Bicep and validates the resulting JSON template.

Automated Validation

The Test-SOFSDeployment.ps1 script validates the full SOFS deployment:

.\tests\Test-SOFSDeployment.ps1 `
    -SOFSAccessPoint "FSLogixSOFS" `
    -ShareNames @("FSLogix") `
    -ClusterName "sofs-cluster" `
    -DomainNetBIOS "IIC"

For Triple layout (three shares):

.\tests\Test-SOFSDeployment.ps1 `
    -SOFSAccessPoint "FSLogixSOFS" `
    -ShareNames @("Profiles", "ODFC", "AppData") `
    -ClusterName "sofs-cluster" `
    -DomainNetBIOS "IIC"

What it checks:

  • SMB share reachability (Test-Path)
  • Share settings: Continuously Available, CachingMode None, correct ScopeName
  • NTFS permissions: CREATOR OWNER, Domain Users, Domain Admins, SYSTEM
  • SMB encryption settings
  • S2D health status
  • Cluster health

Manual Verification

1. SOFS Share Accessibility

From any machine on the compute network:

Test-Path "\\FSLogixSOFS\FSLogix"

"Profiles", "ODFC", "AppData" | ForEach-Object {
    [PSCustomObject]@{ Share = $_; Accessible = (Test-Path "\\FSLogixSOFS\$_") }
}

2. SMB Share Settings

Verify Continuous Availability and caching are correctly configured:

Get-SmbShare -CimSession "sofs-01" -Name "FSLogix" |
    Select-Object Name, ScopeName, ContinuouslyAvailable, CachingMode

Expected:

Property Expected Value
ScopeName SOFS access point name (e.g., FSLogixSOFS)
ContinuouslyAvailable True
CachingMode None

3. Anti-Affinity Verification

Confirm all three SOFS VMs are on separate Azure Local physical nodes:

# Check VM placement
Get-ClusterGroup -Cluster "azl-cluster-01" |
    Where-Object { $_.Name -like "*sofs*" } |
    Select-Object Name, OwnerNode

# Verify affinity rule
Get-ClusterAffinityRule -Name "SOFS-AntiAffinity" -Cluster "azl-cluster-01"

Each VM must show a different OwnerNode.

4. S2D Health

# Storage pool health
Get-StoragePool -CimSession "sofs-cluster" |
    Where-Object { $_.IsPrimordial -eq $false } |
    Select-Object FriendlyName, HealthStatus, OperationalStatus

# Virtual disks
Get-VirtualDisk -CimSession "sofs-cluster" |
    Select-Object FriendlyName, HealthStatus, OperationalStatus, ResiliencySettingName, NumberOfDataCopies

# Physical disks
Get-PhysicalDisk -CimSession "sofs-cluster" |
    Select-Object FriendlyName, HealthStatus, OperationalStatus, Size

All components should report Healthy / OK.

5. Cluster Health

Get-ClusterNode -Cluster "sofs-cluster" | Select-Object Name, State
Get-ClusterGroup -Cluster "sofs-cluster" | Select-Object Name, State, OwnerNode
Get-ClusterQuorum -Cluster "sofs-cluster"

All nodes should be Up. The quorum model should show CloudWitness.

Failover Testing

Test during a maintenance window

Failover testing should be done before production onboarding, not during active user sessions.

Test Procedure

  1. Create a test file on the SOFS share:

    New-Item "\\FSLogixSOFS\FSLogix\test-failover.txt" -Value "Failover test $(Get-Date)"

  2. Identify the current owner of the SOFS role:

    Get-ClusterGroup -Cluster "sofs-cluster" |
    Where-Object { $_.GroupType -eq "ScaleOutFileServer" } |
    Select-Object Name, OwnerNode

  3. Drain a SOFS node (simulates host maintenance or failure):

    # On the Azure Local cluster — drain the host node running one SOFS VM
Suspend-ClusterNode -Name "azl-node-01" -Cluster "azl-cluster-01" -Drain

  4. Verify share remains accessible:

    Test-Path "\\FSLogixSOFS\FSLogix"
Get-Content "\\FSLogixSOFS\FSLogix\test-failover.txt"

  5. Resume the node:

    Resume-ClusterNode -Name "azl-node-01" -Cluster "azl-cluster-01"

What to Expect

  • The share should remain accessible throughout the drain/resume cycle
  • SMB3 transparent failover (via Continuous Availability) handles the reconnection
  • Users with mounted FSLogix profiles experience no disruption
  • S2D continues operating with 2 of 3 nodes — the two-way mirror serves all reads/writes

If Failover Fails

  • Verify ContinuouslyAvailable is $true on all shares
  • Verify the SOFS role is configured as ScaleOutFileServer (not regular FileServer)
  • Check that SMB3 is being used (not SMB2 or earlier)
  • See Troubleshooting for detailed diagnosis

Validation Checklist

  • [ ] All SOFS shares accessible via UNC path
  • [ ] ContinuouslyAvailable = True on all shares
  • [ ] CachingMode = None on all shares
  • [ ] ScopeName matches SOFS access point
  • [ ] Anti-affinity rule active and verified
  • [ ] All SOFS VMs on separate physical nodes (2–16 VMs supported)
  • [ ] S2D pool, virtual disks, and physical disks healthy
  • [ ] Resiliency matches config (NumberOfDataCopies = 2 or 3)
  • [ ] All cluster nodes Up
  • [ ] Cloud witness configured and accessible
  • [ ] NTFS permissions correct (test with a domain user account)
  • [ ] Failover test passed — share accessible during node drain
  • [ ] Test file created and read back after failover
  • [ ] Single layout: single share (FSLogix) present
  • [ ] Triple layout: three shares (Profiles, ODFC, AppData) present
  • [ ] FSRM quotas applied (if configured)
  • [ ] Cloud Cache CCDLocations string matches providers config (if enabled)

Next Steps