virtual machines crash after ReFS unmount - how to avoid the unmount?

Question

occassionally I am observing the following sequence of events:

Microsoft-Windows-Hyper-V-StorageVSP/Admin 22/02/2025 04:59:57

(multiple events of the following type, each relating to a different VM being replicated:)

An I/O request for device 'H:\***.vhdx' took 10043 milliseconds to complete. Operation code = WRITE, Data transfer length = 741376, Status = SRB_STATUS_SUCCESS.

Microsoft-Windows-Storage-Storport/Operational 22/02/2025 05:00:55

(two events)

This is the first instance of the error seen during this time period

on Storport Device (Port = 1, Path = 0, Target = 0, Lun = 0) whose Corresponding Class Disk Device Guid is {91b818fd-fd5e-83d7-c059-ab2794c1d22a}:

The request opcode was 0x35 and completed with SrbStatus 0xE and ScsiStatus 0x0.

The sense code was (0x0,0x0,0x0).

The io latency was 605 ms.

This is the first instance of the error seen during this time period

on Storport Device (Port = 1, Path = 0, Target = 0, Lun = 0) whose Corresponding Class Disk Device Guid is {91b818fd-fd5e-83d7-c059-ab2794c1d22a}:

The request opcode was 0x2A and completed with SrbStatus 0x4 and ScsiStatus 0x0.

The sense code was (0x0,0x0,0x0).

The io latency was 7222 ms.

System/ReFS 22/02/2025 05:00:57

(two events)

The file system was unable to write metadata to the media backing volume H:. A write failed with status "The I/O device reported an I/O error." ReFS will take the volume offline. It may be mounted again automatically.

VolumeId H:

FailureReason 0xc0000185

SourceFile 0x4017

SourceLine 4543

SourceTag 13

SpaceId {00000000-0000-0000-0000-000000000000}

VolumeId H:

FailureReason 0xc0000185

SourceFile 0x402a

SourceLine 4993

SourceTag 20

SpaceId {00000000-0000-0000-0000-000000000000}

at this time, file systems in virtual machines go read-only, and go stuck as they cannot write to disk anymore.

I noticed the writes are slow, but this is on older hardware. I haven´t seen crashes like this on older versions of Windows Server, but am observing this with

Window Server 2025.

Anyone observing similar? Is there a means to wait longer rather than taking volumes offline?

Answer 1

Hello,

Thank you for posting in Microsoft Community forum.

Based on the description, I understand your question is related to vm crash.

1. Put the loading machine in safe mode, and then repeat the operation:

Start your PC in Safe Mode in Windows - Microsoft Support

2. It might also be caused by a faulty driver, try running the Driver Verifier:

Open a command prompt window by selecting Run as administrator and type verifier to open the Driver Verifier Manager.

Driver Verifier - Drivers for Windows | Microsoft Learn

3. Sometimes the problem may not be in the driver, but only in third-party applications and services. Try a clean boot:

How to perform a clean boot on Windows - Microsoft Support

4. Update your drivers, if Driver Verifier has detected a specific one, you can update that driver and see if that fixes the problem.

Press Windows + 'R' and type devmgmt.msc to open Device Manager.

Locate the device and double-click on it to expand.

Double-click on the driver and go to the Drivers tab.

Click "Update Driver" and select "Automatic Search" for drivers.

Windows Client for IT Pros and Windows Server forums are moving to Microsoft Q&A

We’re transitioning to Microsoft Q&A for a more streamlined experience. Starting February 26th, new questions can only be posted on Microsoft Q&A. Existing discussions will remain accessible here. 

Beginning March 3rd customers looking for support on Answers will be automatically redirected to Microsoft Q&A.

Have a nice day. 

Best Regards,

Molly