The essentials
Quick reference
One focused task per row. Jump to the related section for complete, working examples.
| Use | Syntax | Examples |
|---|---|---|
| Check the cluster feature | Get-WindowsFeature -Name Failover-Clustering, `
RSAT-Clustering-PowerShell | View examples |
| Load cluster cmdlets | Import-Module FailoverClusters; Get-Command -Module FailoverClusters |
Select-Object Name | View examples |
| Identify the cluster | Get-Cluster -Name 'AppCluster01' |
Format-List Name, Domain, QuarantineDuration, QuarantineThreshold | View examples |
| Inventory nodes | Get-ClusterNode -Cluster 'AppCluster01' |
Select-Object Name, State, NodeWeight, DynamicWeight, DrainStatus | View examples |
| Inventory clustered roles | Get-ClusterGroup -Cluster 'AppCluster01' |
Select-Object Name, State, OwnerNode, GroupType, Priority | View examples |
| Inspect role resources | Get-ClusterGroup -Cluster 'AppCluster01' -Name 'App Service' |
Get-ClusterResource | View examples |
| Inspect cluster networks | Get-ClusterNetwork -Cluster 'AppCluster01' |
Select-Object Name, State, Role, Address, AddressMask, Metric | View examples |
| Inspect quorum configuration | Get-ClusterQuorum -Cluster 'AppCluster01' |
Format-List Cluster, QuorumResource, QuorumType | View examples |
| List validation tests | Test-Cluster -Cluster 'AppCluster01' -List | View examples |
| Preview cluster validation | Test-Cluster -Cluster 'AppCluster01' -Include `
'Inventory','Network','System Configuration' -WhatIf | View examples |
| Inspect preferred owners | Get-ClusterGroup -Cluster 'AppCluster01' -Name 'App Service' |
Get-ClusterOwnerNode | View examples |
| Preview draining a node | Suspend-ClusterNode -Cluster 'AppCluster01' -Name `
'APPNODE01' -Drain -TargetNode 'APPNODE02' -Wait `
-WhatIf | View examples |
| Move one clustered role | Move-ClusterGroup -Cluster 'AppCluster01' -Name `
'App Service' -Node 'APPNODE02' -Wait 300 | View examples |
| Resume without failback | Resume-ClusterNode -Cluster 'AppCluster01' -Name `
'APPNODE01' -Failback NoFailback | View examples |
| Read recent cluster events | Get-WinEvent -FilterHashtable `
@{LogName='Microsoft-Windows-FailoverClustering/Operational'; `
StartTime=(Get-Date).AddHours(-2)} | View examples |
| Collect focused cluster logs | Get-ClusterLog -Cluster 'AppCluster01' -Node 'APPNODE01' `
-TimeSpan 30 -UseLocalTime -Destination `
'C:/ClusterEvidence' | View examples |
Failover clustering coordinates workload ownership, health detection, recovery, and quorum across Windows Server nodes; it does not make an application interruption-proof by itself. Establish a validated, supported design first, then resolve the exact cluster, node, and role before every change. Run mutations elevated under an approved maintenance plan, preserve quorum and management access, and verify workload health from the client path after any movement.
Step by step
Detailed examples
Establish supported scope and administrative prerequisites
Failover Clustering is a Windows Server feature supported on Standard and Datacenter editions, while particular clustered roles can add edition, licensing, hardware, storage, or networking requirements. Install Failover-Clustering on every node and RSAT-Clustering-PowerShell on each management system; feature installation itself normally does not require a restart, but servicing or dependent roles might. Production nodes should run the same Windows Server version and supported hardware should pass cluster validation. Traditional clusters normally use nodes in one Active Directory domain and require rights to create or use the cluster name object; workgroup and AD-detached clusters are specialized designs with different constraints, not drop-in exceptions. Use an elevated Windows PowerShell session with local administrative or delegated cluster rights. The -Cluster parameter performs cluster-aware remote management, but individual commands can impose extra remoting constraints; notably Get-ClusterLog cannot run remotely without CredSSP, so collect it locally or through an approved management path.
$identity = [Security.Principal.WindowsIdentity]::GetCurrent()
$principal = [Security.Principal.WindowsPrincipal]::new($identity)
[pscustomobject]@{
ComputerName = $env:COMPUTERNAME
Edition = (Get-ComputerInfo -Property WindowsProductName).WindowsProductName
Elevated = $principal.IsInRole([Security.Principal.WindowsBuiltInRole]::Administrator)
}
Get-WindowsFeature -Name Failover-Clustering, RSAT-Clustering-PowerShell
Import-Module FailoverClusters
Get-Command -Module FailoverClusters | Select-Object NameNote: Install-WindowsFeature changes every named server and can trigger a pending restart through dependencies or later patching. Check InstallState and RestartNeeded under a maintenance plan instead of installing features from discovery scripts.
Inventory ownership, resources, and communication paths
Cluster health is a relationship among nodes, roles, resources, networks, and workload-specific dependencies. Query the named cluster rather than relying on the local default, record node and role ownership before maintenance, and drill into resources when a role is degraded. Network Role controls whether a network carries cluster traffic, client traffic, both, or neither; changing it can sever heartbeats or client access. Read-only cmdlets do not prove application health, so pair the inventory with service-specific probes from outside the cluster.
$clusterName = 'AppCluster01'
Get-Cluster -Name $clusterName | Format-List Name, Domain, QuarantineDuration, QuarantineThreshold
Get-ClusterNode -Cluster $clusterName | Select-Object Name, State, NodeWeight, DynamicWeight, DrainStatus
Get-ClusterGroup -Cluster $clusterName | Select-Object Name, State, OwnerNode, GroupType, Priority
Get-ClusterGroup -Cluster $clusterName -Name 'App Service' | Get-ClusterResource | Select-Object Name, ResourceType, State, OwnerGroup, OwnerNode
Get-ClusterNetwork -Cluster $clusterName | Select-Object Name, State, Role, Address, AddressMask, MetricProtect quorum and schedule validation deliberately
Quorum prevents two partitions from accepting writes at the same time; it is not simply a node-count setting. Inspect the existing witness, vote state, site topology, and failure model before any quorum change. Set-ClusterQuorum has no WhatIf parameter, so a witness change needs an approved design, verified credentials or share permissions, and a recovery plan. Test-Cluster can run before or after creation, but tests are not universally passive: storage validation can take eligible storage offline, -Force can take specified clustered disks or pools offline, and some CSV tests temporarily create a local CliTest2 account. Use -List and -WhatIf for planning, then schedule the real supported validation set during a suitable window and review its report.
$clusterName = 'AppCluster01'
Get-ClusterQuorum -Cluster $clusterName | Format-List Cluster, QuorumResource, QuorumType
Get-ClusterNode -Cluster $clusterName | Select-Object Name, State, NodeWeight, DynamicWeight
Test-Cluster -Cluster $clusterName -List
Test-Cluster -Cluster $clusterName -Include 'Inventory','Network','System Configuration' -WhatIfNote: WhatIf confirms command intent, not hardware support, workload compatibility, witness reachability, report success, or the runtime impact of the eventual validation. Never add -Force to exploratory storage validation.
Drain, maintain, and resume one node at a time
Before maintenance, confirm quorum headroom, destination capacity, possible and preferred owners, replication health, and application-specific failover behavior. Suspend-ClusterNode -Drain uses cluster placement logic and supports WhatIf; -ForceDrain may stop workloads that cannot move safely and should not be a routine shortcut. Move-ClusterGroup and Resume-ClusterNode do not support WhatIf, and a successful move can still interrupt connections or expose application faults. Use explicit role and node names, monitor completion, validate the workload externally, and resume with NoFailback unless an approved policy calls for movement back. Operating-system or firmware servicing can require a reboot even though draining and resuming do not.
$clusterName = 'AppCluster01'
$sourceNode = 'APPNODE01'
$targetNode = 'APPNODE02'
$roleName = 'App Service'
Get-ClusterGroup -Cluster $clusterName -Name $roleName | Get-ClusterOwnerNode
Suspend-ClusterNode -Cluster $clusterName -Name $sourceNode -Drain -TargetNode $targetNode -Wait -WhatIf
# After approval: Move-ClusterGroup -Cluster $clusterName -Name $roleName -Node $targetNode -Wait 300
# After maintenance and workload checks: Resume-ClusterNode -Cluster $clusterName -Name $sourceNode -Failback NoFailbackNote: The commented Move-ClusterGroup and Resume-ClusterNode calls deliberately omit WhatIf because those cmdlets do not implement it. Execute them only through a reviewed runbook with current capacity and quorum evidence.
Collect time-bounded evidence before changing state
Correlate the FailoverClustering operational log, cluster log, resource states, application logs, storage and network telemetry, and exact UTC or local timestamps before restarting resources. Get-ClusterLog creates files and can be expensive across all nodes, so constrain the node and timespan and use a protected destination with adequate space. It uses GMT by default; -UseLocalTime improves correlation only when the incident record states the timezone. Because remote collection has CredSSP restrictions, prefer running it on a cluster node or a secured administrative session rather than weakening authentication. Log collection is evidence, not a repair.
$since = (Get-Date).AddHours(-2)
Get-WinEvent -FilterHashtable @{
LogName = 'Microsoft-Windows-FailoverClustering/Operational'
StartTime = $since
} | Select-Object TimeCreated, Id, LevelDisplayName, Message
Get-ClusterLog -Cluster 'AppCluster01' -Node 'APPNODE01' -TimeSpan 30 -UseLocalTime -Destination 'C:\ClusterEvidence'Note: The destination is a placeholder. Create and ACL an incident-specific directory first, and avoid collecting secrets or unrelated tenant data into broadly accessible paths.
Sources and further reading
References
Authoritative documentation used to verify and expand this cheat sheet.
Help us improve
Found a typo or missing example?
Tell us what would make this cheat sheet clearer, more complete, or more useful.