Share via

Windows Failover Cluster IO - Role vs Disk Owner

Eric 0 Reputation points
2025-05-08T13:12:38.1233333+00:00

We have a Windows Server 2022 cluster using iscsi mounted shared disks and running Hyper-V VM's. They have a direct connection with no gateway to the storage array on one nic and the company network on the other nic. When in Failover Cluster Manager, if one node has ownership of a running VM, but another node has ownership of the CSV disk it's on, I'm noticing that when I test a file copy from within the VM, the data isn't written directly to the iscsi disk on the VM node owner. It's transfered to the disk owner node over the network, then written to disk. If I change the owner of the disk to the same node running the VM, it writes directly to the disk. It's documented that it shouldn't matter which node owns a CSV disk... any node should be able to write directly to the disk if necessary. Any ideas why this might be happening?

Windows Server High availability Clustering and high availability

1 answer

Sort by: Most helpful
Most helpful Newest Oldest
  1. Henry Mai 795 Reputation points
    2025-05-09T09:59:32.24+00:00

    Hi, I am Henry and I'd like to help.

    It sounds like you encountered an unexpected I/O redirection issue in your Windows Server 2022 Failover Cluster setup . Normally, Cluster Shared Volumes (CSV) allow any node to write directly to the disk, regardless of ownership. However, in some cases, I/O redirection can occur, causing data to be transferred over the network instead of being written directly.

    1. CSV Disk in Redirected Mode
    • Even though CSV disks should allow direct writes, sometimes they enter Redirected Mode, forcing I/O to go through the disk owner.
    • You can check the status using PowerShell: Get-ClusterSharedVolumeState
    • If the disk is in Redirected Mode, try restarting the cluster service or manually failing over the disk.
    1. Failover Cluster Communication Issues
    • Since your setup involves iSCSI-mounted shared disks, ensure that all nodes have proper access to the storage array.
    • Check network configurations to confirm that the storage NICs are correctly isolated from the company network.
    1. CSV Coordinator Node Behavior
    • Windows Failover Clustering assigns a Coordinator Node for CSV management.
    • If the VM is running on a different node than the disk owner, the cluster may temporarily redirect I/O through the disk owner.
    • Failing over the role and disk (as you did) likely reset the coordinator node, resolving the issue.
    1. Check Cluster Logs for Errors
    • Run the following command to check for failover cluster errors: Get-ClusterLog -TimeSpan 5 -Destination C:\ClusterLogs
    • Look for any I/O redirection warnings or network-related errors.
    1. Update Windows Server & Storage Drivers
    • Ensure that Windows Server 2022 is fully updated.
    • Check for firmware updates on your storage array and NIC drivers.

    Let me know if you need further assistance

    0 comments

Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.