Get-S2DCapacityWaterfall¶

Computes the complete 7-stage capacity accounting pipeline from raw physical disk capacity through resiliency overhead to final usable VM space.

Syntax¶

Get-S2DCapacityWaterfall

No parameters. All inputs are sourced from the module session cache — the function automatically runs prerequisite collectors if their data is not already cached.

Prerequisites¶

Get-S2DCapacityWaterfall depends on data from three other collectors. It runs them automatically if needed:

Get-S2DPhysicalDiskInventory — capacity disk sizes, largest drive, node count
Get-S2DStoragePoolInfo — pool total size and free space
Get-S2DVolumeMap — volume footprints and infrastructure volume classification

To run collectors explicitly in advance:

Connect-S2DCluster -ClusterName "c01-prd-bal" -Credential $cred
Get-S2DPhysicalDiskInventory | Out-Null
Get-S2DStoragePoolInfo       | Out-Null
Get-S2DVolumeMap             | Out-Null
$wf = Get-S2DCapacityWaterfall

Output¶

Returns S2DCapacityWaterfall — a single object with top-level summary properties and a Stages array.

Top-level properties¶

Property	Type	Description
`Stages`	`S2DWaterfallStage[]`	The 7 pipeline stages (see below)
`RawCapacity`	`S2DCapacity`	Stage 1 total (sum of capacity-tier disk bytes)
`UsableCapacity`	`S2DCapacity`	Stage 7 total (final usable capacity)
`ReserveRecommended`	`S2DCapacity`	Recommended reserve: `min(NodeCount, 4) × LargestDrive`
`ReserveActual`	`S2DCapacity`	Actual free pool space
`ReserveStatus`	`string`	`Adequate`, `Warning`, or `Critical`
`IsOvercommitted`	`bool`	Whether `OvercommitRatio > 1.0`
`OvercommitRatio`	`double`	From `StoragePool.OvercommitRatio`
`NodeCount`	`int`	Node count used in reserve calculation
`BlendedEfficiencyPercent`	`double`	Average resiliency efficiency across workload volumes

Stage object properties¶

Property	Type	Description
`Stage`	`int`	Stage number (1–7)
`Name`	`string`	Stage label
`Size`	`S2DCapacity`	Remaining capacity at this stage
`Delta`	`S2DCapacity`	Bytes consumed vs previous stage (`$null` if none)
`Description`	`string`	Human-readable explanation of what changed
`Status`	`string`	`OK` — all stages are always OK; reserve health is on `ReserveStatus` only

The 7 Stages¶

flowchart LR
    S1["Stage 1\nRaw Physical"] --> S2["Stage 2\nVendor Label"]
    S2 --> S3["Stage 3\nPool Overhead"]
    S3 --> S4["Stage 4\nReserve"]
    S4 --> S5["Stage 5\nInfra Volume"]
    S5 --> S6["Stage 6\nAvailable"]
    S6 --> S7["Stage 7\nUsable Capacity"]

Stage	Name	What changes
1	Raw Physical	Sum of all capacity-tier disk bytes (cache disks excluded)
2	Vendor Label (TB)	Display stage: shows vendor TB label vs Windows TiB. No bytes change.
3	Pool (after overhead)	Actual pool total size — captures ~0.5–1% metadata overhead
4	After Reserve	Subtracts `min(NodeCount, 4) × LargestCapacityDrive`
5	After Infra Volume	Subtracts Azure Local infrastructure volume footprint
6	Available	Pool space available for workload volumes
7	Usable Capacity	Stage 6 ÷ resiliency factor (e.g. ÷3 for three-way mirror). Pipeline terminus.

See Capacity Math for a full explanation of each stage.

Reserve Status¶

Stage 4 computes a reserve status based on how actual free pool space compares to the recommended reserve:

Status	Condition
`Adequate`	`ReserveActual ≥ ReserveRecommended`
`Warning`	`ReserveActual ≥ 50% of ReserveRecommended`
`Critical`	`ReserveActual < 50% of ReserveRecommended`

The reserve recommendation follows the Microsoft S2D guidance: keep min(NodeCount, 4) capacity drive equivalents unallocated so the pool can rebuild after a drive or node failure.

Critical reserve status

A Critical reserve means the pool cannot guarantee a complete rebuild after the loss of the largest drive. Immediately free pool space by shrinking or removing volumes.

Session Behavior¶

Results are cached in $Script:S2DSession.CollectedData['CapacityWaterfall']. Calling Get-S2DCapacityWaterfall twice only queries the cluster once.

Examples¶

# Full waterfall pipeline
$wf = Get-S2DCapacityWaterfall
$wf.Stages | Format-Table Stage, Name, Size, Delta, Status

# Summary
"Raw:        $($wf.RawCapacity.Display)"
"Usable:     $($wf.UsableCapacity.Display)"
"Reserve:    $($wf.ReserveStatus) — Actual: $($wf.ReserveActual.Display), Recommended: $($wf.ReserveRecommended.Display)"
"Efficiency: $($wf.BlendedEfficiencyPercent)%"

# Check reserve
if ($wf.ReserveStatus -ne 'Adequate') {
    Write-Warning "Reserve is $($wf.ReserveStatus): actual $($wf.ReserveActual.Display) vs recommended $($wf.ReserveRecommended.Display)"
}

# Stage 4 only
$wf.Stages | Where-Object Stage -eq 4 | Format-List

Troubleshooting¶

Stage 7 (Usable Capacity) is 0 TiB

Usable capacity is zero when no workload volumes exist, or all volumes are classified as infrastructure. Check:

Get-S2DVolumeMap | Format-Table FriendlyName, IsInfrastructureVolume, Size

Stage 2 shows no change

Stage 2 is informational only — it shows the vendor-labeled TB value alongside the Windows TiB value. The Size property does not change from Stage 1 to Stage 2 because the underlying bytes are the same. See TiB vs TB for an explanation of the discrepancy.

NodeCount affects reserve calculation

The reserve uses min(NodeCount, 4) where NodeCount comes from $Script:S2DSession.Nodes.Count. If you call Get-S2DCapacityWaterfall without an active Connect-S2DCluster session (e.g., passing -CimSession directly to collectors), the node count falls back to the disk symmetry group count or 4. Use Connect-S2DCluster for accurate node counts.