Get-S2DPhysicalDiskInventory¶

Inventories all physical disks in the S2D cluster with health, capacity, wear, and reliability data — collected per node over WinRM or local execution.

Syntax¶

Get-S2DPhysicalDiskInventory [[-NodeName] <string[]>] [[-CimSession] <CimSession>]

Parameters¶

Parameter	Type	Description
`-NodeName`	`string[]`	Limit results to one or more specific node names.
`-CimSession`	`CimSession`	Override the module session CimSession. Useful for ad-hoc calls without a full `Connect-S2DCluster` session.

Output¶

Returns PSCustomObject[] — one object per physical disk.

Property	Type	Description
`NodeName`	`string`	Node that reported this disk
`DiskNumber`	`int`	Windows disk number (from `Get-Disk`)
`UniqueId`	`string`	Storage subsystem unique identifier
`FriendlyName`	`string`	Vendor-reported model string
`SerialNumber`	`string`	Drive serial number
`Model`	`string`	Disk model identifier
`MediaType`	`string`	`NVMe`, `SSD`, `HDD`, or `SCM`
`BusType`	`string`	`NVMe`, `SATA`, `SAS`, `RAID`, etc.
`FirmwareVersion`	`string`	Active firmware version string
`Manufacturer`	`string`	Drive vendor
`Role`	`string`	`Cache`, `Capacity`, or `Unknown`
`Usage`	`string`	S2D usage classification: `Auto-Select`, `Journal`, `Retired`
`CanPool`	`bool`	Whether the disk is eligible for pooling
`IsPoolMember`	`bool`	Whether this disk is currently a member of the S2D storage pool. Use this to filter out boot drives (BOSS) and SAN-presented LUNs that happen to be visible to the node.
`HealthStatus`	`string`	`Healthy`, `Warning`, or `Unhealthy`
`OperationalStatus`	`string`	Operational state from WMI
`PhysicalLocation`	`string`	Physical bay/slot location string (when available)
`SlotNumber`	`int`	Numeric slot number (when available)
`Size`	`S2DCapacity`	Disk capacity (TiB + TB dual-display)
`SizeBytes`	`int64`	Raw capacity in bytes
`Temperature`	`int`	Drive temperature in Celsius
`WearPercentage`	`int`	NVMe wear percentage (0–100)
`PowerOnHours`	`int64`	Total power-on hours
`ReadErrors`	`int64`	Uncorrected read errors
`WriteErrors`	`int64`	Uncorrected write errors
`ReadLatency`	`int64`	Average read latency (when available)
`WriteLatency`	`int64`	Average write latency (when available)

Pool Membership Filter¶

Get-S2DPhysicalDiskInventory returns every disk visible to every node — boot drives (Dell BOSS, HPE SmartArray M.2, etc.), SAN-presented LUNs, and anything else Windows can see. This is intentional: the collector gives downstream consumers full fidelity, and the JSON / CSV exports always preserve every disk.

The rendered reports (HTML, Word, PDF, Excel Physical Disk Inventory table) filter to pool members only by default. The filter is driven by the IsPoolMember boolean on each disk:

IsPoolMember = $true — disk is a member of the S2D storage pool. Included.
IsPoolMember = $false — disk is visible to the node but not a pool member (boot drive, SAN LUN, spare). Excluded from the rendered table.

To include non-pool disks in rendered reports:

Invoke-S2DCartographer -ClusterName <name> -IncludeNonPoolDisks
New-S2DReport -InputObject $data -Format All -IncludeNonPoolDisks

Health checks (DiskSymmetry, DiskHealth, NVMeWear, FirmwareConsistency, RebuildCapacity) always operate on pool members only, regardless of this switch — non-pool disks are not in S2D scope.

Role Classification¶

S2DCartographer classifies each disk as Cache, Capacity, or Unknown based on two factors:

Usage field — disks with Usage = Journal are always Cache.
Media type heuristic — within the pool, disks with the highest media rank (NVMe > SSD > HDD) are classified as Cache; lower-rank disks are Capacity.

On all-NVMe clusters where all drives share the same media type, all pool disks are classified as Capacity — S2D uses a software write-back cache in this configuration rather than a physical cache tier.

Anomaly Detection¶

On first collection, Get-S2DPhysicalDiskInventory emits Write-Warning output for any of the following conditions:

Disk symmetry — disk count differs across nodes (asymmetric configuration is unsupported by S2D)
Mixed capacity sizes — capacity-tier disks of different sizes within the cluster
Firmware inconsistency — disks of the same model running different firmware versions
Non-healthy disks — any disk not in Healthy state

These warnings surface immediately in the console and are also detected formally by Get-S2DHealthStatus as DiskSymmetry, FirmwareConsistency, and DiskHealth checks.

CIM Sources¶

Get-PhysicalDisk                           # per node
Get-PhysicalDisk | Get-StorageReliabilityCounter  # per disk
Get-Disk                                   # enriches physical location and slot
Get-StoragePool | Get-PhysicalDisk         # pool membership check

All queries run per node when PSSession is available. When only a CimSession is present, the function creates temporary per-node CIM sessions from the node list stored in $Script:S2DSession.Nodes.

Session Behavior¶

Connection type	How disks are collected
PSSession (standard remote)	`Invoke-Command` to each node; runs locally on that node
CimSession only	Creates temporary per-node `CimSession` from the node list
`-Local` switch	Runs against `$env:COMPUTERNAME` directly
`-NodeName` filter	Applied after collection; no impact on CIM sessions opened

Results are cached in $Script:S2DSession.CollectedData['PhysicalDisks'] after first collection. Subsequent callers (including Get-S2DCapacityWaterfall, Get-S2DHealthStatus, and Get-S2DCacheTierInfo) read from cache without a network round-trip.

Examples¶

# Full inventory — all nodes
Get-S2DPhysicalDiskInventory | Format-Table NodeName, FriendlyName, Role, Size, WearPercentage, HealthStatus

# NVMe wear report — highlight drives over 60%
Get-S2DPhysicalDiskInventory |
    Where-Object { $_.MediaType -eq 'NVMe' } |
    Sort-Object WearPercentage -Descending |
    Format-Table NodeName, FriendlyName, WearPercentage, PowerOnHours

# Single-node inventory
Get-S2DPhysicalDiskInventory -NodeName "node01"

# Capacity-tier disks only
Get-S2DPhysicalDiskInventory | Where-Object Role -eq 'Capacity' | Format-Table FriendlyName, Size, FirmwareVersion

Troubleshooting¶

Returns empty or partial results

If the command returns nothing or only disks from some nodes, at least one node is unreachable. Use -Verbose to see per-node connection attempts:

Get-S2DPhysicalDiskInventory -Verbose

Look for WARNING: Could not collect disks from node lines.

Reliability counters may be null

WearPercentage, PowerOnHours, ReadErrors, and WriteErrors are sourced from Get-StorageReliabilityCounter. Some drivers and firmware versions do not expose all reliability counters — these properties will be $null when the counter is unavailable.

Running on a cluster node

If you run Connect-S2DCluster -Local directly on a node, the function skips per-node session creation and queries the local host only. All pool disks visible to that node will be returned, but disks exclusively visible to other nodes may be missing.