Hello,
Since you have already cleared the cluster, I would focus on identifying what happens during CSV creation and node maintenance before making further changes to the storage configuration.
Even with updated firmware, drivers, and Windows updates, S2D instability can still be caused by network configuration differences, unsupported storage controllers, disk health issues, or leftover storage metadata.
1. Capture the event details
Open Event Viewer (eventvwr.msc) and locate Events 1101 and 1108 immediately after reproducing the issue.
Check the Event Source, complete General message, timestamp, and Details tab, including the XML data.
You can also capture the events using PowerShell:
$Start = Get-Date
# Reproduce the CSV creation or maintenance operation here.
Get-WinEvent -FilterHashtable @{
LogName = 'System'
StartTime = $Start
} | Where-Object { $_.Id -in 1101,1108 } |
Select-Object TimeCreated, ProviderName, Id, LevelDisplayName, Message |
Format-List
The event source and complete messages are essential because event IDs alone do not identify the failing component.
2. Check storage health and disk availability
Run:
Get-PhysicalDisk | Format-Table FriendlyName, OperationalStatus, HealthStatus, CanPool, Usage
Get-VirtualDisk | Format-Table FriendlyName, OperationalStatus, HealthStatus
Get-StoragePool -IsPrimordial $false | Format-Table FriendlyName, OperationalStatus, HealthStatus
Get-StorageJob
Check for unhealthy or degraded disks, incomplete storage jobs, and disks that cannot be pooled. If CanPool is False, investigate the reason before attempting to clean or reset any disk.
3. Verify network and RDMA configuration
S2D depends heavily on reliable communication between nodes. Check that the network adapter configuration is consistent across all four nodes, particularly RDMA settings and, if using RoCE, DCB, PFC and ETS configuration.
Run:
Get-NetAdapterRdma
Get-SmbMultichannelConnection
Get-NetAdapter | Format-Table Name, Status, LinkSpeed
These commands report information from the local node, so run them on all four nodes and compare the results. Look for unexpected differences, unavailable RDMA interfaces, or missing SMB Multichannel connections.
4. Verify storage controller compatibility
Confirm that the storage controllers are configured in a supported HBA or pass-through mode rather than traditional hardware RAID mode.
Also verify that the exact server, controller, drive models and firmware combinations are supported for Storage Spaces Direct. Having the latest firmware does not necessarily mean the hardware configuration is supported.
5. Run cluster validation
Since the cluster has already been cleared, run cluster validation on the current configuration and review any storage or network-related warnings before proceeding with additional configuration changes.
Test-Cluster -Include "Storage Spaces Direct","Inventory","Network","System Configuration"
Review the storage and network results carefully, particularly any warnings or failures relating to disk eligibility, connectivity, and configuration consistency.
6. Collect cluster logs
Immediately after reproducing the issue, collect the cluster logs:
Get-ClusterLog -UseLocalTime -TimeSpan 30 -Destination C:\ClusterLogs
This collects the last 30 minutes of cluster activity. Review the logs from all four nodes and correlate the entries with the timestamps of Events 1101 and 1108.
Microsoft documentation for further guidance:
Troubleshoot Storage Spaces Direct
Storage Spaces Direct hardware requirements
Troubleshoot storage issues in Hyper-V and Windows Server failover clusters
If the physical disks and virtual disks report healthy states but the events continue, the event provider and complete messages, together with the cluster logs and validation report, will be important for identifying the underlying failure.
If this answer helps resolve your issue, please consider accepting it as the answer. This will also help other community members facing the same problem find the solution more easily.