Skip to content

What-If Capacity Modeling

Invoke-S2DCapacityWhatIf models the capacity impact of proposed hardware changes — adding nodes, adding disks, replacing disks, or changing resiliency — without touching the live cluster.

It applies scenario modifications to a baseline capacity waterfall and recomputes all 7 stages, producing a side-by-side before/after comparison with per-stage deltas.

Syntax

# From a JSON snapshot
Invoke-S2DCapacityWhatIf -BaselineSnapshot <path> [scenario parameters] [-OutputDirectory <path>] [-Format <Html|Json>] [-PassThru]

# From a piped S2DClusterData object
Invoke-S2DCartographer -ClusterName <name> -PassThru |
    Invoke-S2DCapacityWhatIf [scenario parameters]

# Against a live cluster (collects then models — does not re-hit the cluster on repeated runs)
Invoke-S2DCapacityWhatIf -ClusterName <name> [scenario parameters]

Parameters

Baseline input (choose one)

Parameter Description
-BaselineSnapshot Path to a JSON snapshot produced by Invoke-S2DCartographer (SchemaVersion 1.0).
-InputObject An S2DClusterData object piped from Invoke-S2DCartographer -PassThru.
-ClusterName Connect to a live cluster, collect data, then model. Does not re-collect on repeated -PassThru calls.

Scenario modifications (combinable)

Parameter Description
-AddNodes <int> Number of nodes to add. New nodes are assumed to have the same disk configuration as existing nodes (average disk count × disk size).
-AddDisksPerNode <int> Number of capacity disks to add per node. Enforces symmetry — all existing nodes receive the same count.
-NewDiskSizeTB <double> Size in decimal TB for new disks added via -AddNodes or -AddDisksPerNode. Defaults to the existing largest disk.
-ReplaceDiskSizeTB <double> Model replacing all capacity disks with disks of this size in decimal TB. Node count and disk count remain the same.
-ChangeResiliency <int> Override the resiliency factor (NumberOfDataCopies). 2 = 2-way mirror, 3 = 3-way mirror.

Scenario parameters can be combined in a single invocation for a composite what-if.

Output

Parameter Description
-OutputDirectory <string> Directory to write reports to. If omitted, no files are written.
-Format <string[]> Report formats: Html (default), Json, or both.
-PassThru Return the S2DWhatIfResult object to the pipeline.

Output object — S2DWhatIfResult

S2DWhatIfResult
├── ScenarioLabel          string          Human-readable description of applied changes
├── BaselineNodeCount      int             Node count before the scenario
├── ProjectedNodeCount     int             Node count after the scenario
├── DeltaUsableTiB         double          Net change in usable capacity (TiB)
├── DeltaUsableTB          double          Net change in usable capacity (TB)
├── GeneratedAt            string          ISO 8601 timestamp
├── BaselineWaterfall      S2DCapacityWaterfall
├── ProjectedWaterfall     S2DCapacityWaterfall
└── DeltaStages[]
    ├── Stage              int             Stage number (1–8)
    ├── Name               string          Stage name
    ├── BaselineTiB        double          Baseline size at this stage
    ├── ProjectedTiB       double          Projected size at this stage
    ├── DeltaTiB           double          Difference (positive = gain)
    ├── BaselineTB         double
    ├── ProjectedTB        double
    └── DeltaTB            double

Examples

Add 2 nodes

Model scaling out from 4 to 6 nodes on a cluster with 4 × 3.84 TB drives per node:

Invoke-S2DCapacityWhatIf -BaselineSnapshot C:\snapshots\clus01-20260101.json -AddNodes 2

The projected waterfall recomputes all 7 stages with 6 nodes × 4 disks. Reserve deduction increases (min(6,4) × largest drive remains capped at 4), but raw capacity grows proportionally.

Add 4 disks per node (same disk size)

Invoke-S2DCapacityWhatIf `
    -BaselineSnapshot C:\snapshots\clus01-20260101.json `
    -AddDisksPerNode 4

Disk count per node grows from 4 to 8. The new disks are assumed to be the same size as the existing largest disk. Raw bytes and all downstream stages increase proportionally.

Add disks with a specific size

When adding drives of a different size (e.g., upgrading to 7.68 TB NVMe):

Invoke-S2DCapacityWhatIf `
    -BaselineSnapshot C:\snapshots\clus01-20260101.json `
    -AddDisksPerNode 4 `
    -NewDiskSizeTB 7.68

The projected largest-disk size used for reserve calculation will be max(existing, new).

Replace all capacity disks

Model replacing 3.84 TB drives with 7.68 TB drives across all nodes (same node and disk count):

Invoke-S2DCapacityWhatIf `
    -BaselineSnapshot C:\snapshots\clus01-20260101.json `
    -ReplaceDiskSizeTB 7.68

Total disk count is unchanged. All stages recompute using the new drive size as both the per-disk size and the largest-disk reserve multiplier.

Change resiliency (3-way → 2-way mirror)

Invoke-S2DCapacityWhatIf `
    -BaselineSnapshot C:\snapshots\clus01-20260101.json `
    -ChangeResiliency 2

Stage 7 (Usable Capacity) changes. All earlier stages are unchanged because resiliency does not affect pool or reserve math. Blended efficiency changes from 33.3% to 50.0%.

Resiliency change is a cluster-level operation

Changing resiliency on a live cluster requires draining, stopping, and reconfiguring volumes. This cmdlet models the capacity outcome only — it does not make any live changes.

Composite: add 2 nodes + more disks + new resiliency

Invoke-S2DCapacityWhatIf `
    -BaselineSnapshot C:\snapshots\clus01-20260101.json `
    -AddNodes 2 `
    -AddDisksPerNode 4 `
    -NewDiskSizeTB 7.68 `
    -ChangeResiliency 2

All four modifications are applied together. The ScenarioLabel property lists each change: +2 nodes, +4 disks/node, Resiliency → 2-way mirror.

Pipeline from live run

Run the full audit on a cluster and immediately feed the result into what-if modeling:

$data = Invoke-S2DCartographer -ClusterName clus01 -PassThru
$whatif = $data | Invoke-S2DCapacityWhatIf -AddNodes 2 -PassThru
$whatif.DeltaUsableTiB

Generate HTML and JSON reports

Invoke-S2DCapacityWhatIf `
    -BaselineSnapshot C:\snapshots\clus01-20260101.json `
    -AddNodes 2 `
    -AddDisksPerNode 4 `
    -OutputDirectory C:\Reports\WhatIf `
    -Format Html, Json

Writes two files to C:\Reports\WhatIf\: - whatif-20260101-1430.html — interactive before/after waterfall chart with KPI summary and delta table - whatif-20260101-1430.json — structured JSON with full baseline and projected waterfall data

How the model works

Baseline extraction

When loading from a JSON snapshot, the cmdlet extracts:

  • Node count — from NodeCount field, or counted from unique node names on capacity disks
  • Disks per node — total capacity disks ÷ node count (averaged, rounded)
  • Largest disk size — maximum SizeBytes across pool-member capacity-tier disks
  • Pool total / free bytes — from StoragePool.TotalSize and RemainingSize
  • Infrastructure volume bytes — from FootprintOnPool (or Size) of volumes where IsInfrastructureVolume = true
  • Resiliency factor — from StoragePool.ResiliencySettings Mirror entry NumberOfDataCopies; defaults to 3.0

Projection calculation

Scenario parameters modify the extracted baseline inputs:

  1. -ReplaceDiskSizeTB — overrides the per-disk size used for raw byte calculation
  2. -NewDiskSizeTB — sets the size for newly added disks; projected largest-disk = max(existing, new)
  3. -AddDisksPerNode — increments disks-per-node count
  4. -AddNodes — increments node count
  5. -ChangeResiliency — overrides the resiliency factor

Projected raw bytes = projected node count × projected disks per node × projected disk size

Projected pool = estimated as projected raw × 99% (standard pool overhead).

Infrastructure volume bytes are unchanged — the model assumes existing infra volumes are unchanged regardless of hardware changes.

The same Invoke-S2DWaterfallCalculation pure function computes both the baseline and projected waterfall from these inputs.

Limitations

  • New nodes are assumed to be symmetric — same disk count and size as the model inputs. If your real nodes will have a different configuration, chain -AddNodes and -ReplaceDiskSizeTB together.
  • The model does not account for cluster rebuild time, performance impact, or rolling upgrade constraints.
  • Thin-provisioned volume growth risk is modeled by Get-S2DHealthStatus (checks 6 and 11) — not by this cmdlet.