Important: This issue does not occur when copying FROM the PC to the SoftRAID drive.

On Mac (using Sonoma 14.5) the Softraid volume works perfectly read/write.

System Specifications

• OWC Express 4M2
  • Using 4x Western Digital WD_BLACK 2TB SN770 NVMe
• PC Motherboard: ASUS ProArt X670E-CREATOR WIFI
  • Manual
• RAM: 32GB
• GPU: NVIDIA RTX 4080
• Thunderbolt: 2x Thunderbolt ports on motherboard
• Mac Studio: with M2 Ultra / 64GB Ram

Software Environment

• SoftRAID Version: 8 (Also tested with version 4 / same issue)
• Windows Versions Tested: Windows 10 and Windows 11 (same issue on both)
• RAID Format: APFS and HFS+ (issue occurs with both)

Specific Scenario

• Copy of a 425gb folder: Image sequence folder containing 13,674 items, totaling 425.58 GB
• Initial Transfer Speed: Starts at 1.9 GB/s for about 40% of the transfer
• Issue Onset: Speed suddenly drops to 30 MB/s, system becomes unresponsive
• Outcome: Unable to shut down normally, requires hard reset
• Additional Impact: Sometimes compromises RAID integrity, preventing mounting on Mac after the incident

Troubleshooting Steps Taken

1. BIOS Updates:
   • Tested latest BIOS version.
   • Reverted to oldest version.
   • Tried a version in between..
   • Result: No difference observed
2. Thunderbolt Settings in BIOS:
   • Thunderbolt/USB4 Support:
   • Security Level:
   • MMIO Resource:
   • Wake Up Command:
   • Result: Adjusting these settings did not resolve the issue
3. Thunderbolt Driver (PC side):
   • Tested NVM firmware versions 31 through 38
   • Result: No improvement
4. Thunderbolt Ports:
   • Tested both Thunderbolt ports on the motherboard
   • Result: Issue persists on both ports

Additional Observations

1. Mac Performance:
   • RAID drive works perfectly on Mac, with read/write speeds around 2300 MB/s (tested with Blackmagic Disk Speed Test)
2. Windows Performance:
   • Initial Benchmark: 950 MB/s read, 2300 MB/s write
   • Then after changing SoftRAID volume 'performance settings' from 'Better Performance' to 'quick removal' and back to 'Better Performance' I get 2300 MB/s read, 2300 MB/s write in Blackmagic Disk Speed Test.
3. Windows Event Viewer:
   • Warning observed: "The driver detected that the device \Device\Harddisk5\DR5 has its write cache enabled. Data Corruption may occur."
   • This warning appears for each drive in the SoftRAID volume
4. Post-Issue Behavior:
   • After a crash during copy, the SoftRAID volume is unmounted in SoftRAID and cannot be re-mounted.

Questions for SoftRAID Team

1. Do I need to do anything special to prepare the Express 4M2 to work on windows/mac? Like preparing each NVME a specific way?
2. Are there known compatibility issues between SoftRAID 8 and Windows 10/11 on systems with Thunderbolt 4?
3. Could the write cache warning in Event Viewer be related to this issue? If so, how can it be addressed?
4. Are there any recommended settings or configurations for high-performance setups like this one?
5. Is there a way to improve the stability of large file transfers from the SoftRAID volume to Windows?
6. Are there any diagnostic tools or logs that could provide more insight into what's happening during these crashes?

Any assistance or direction would be greatly appreciated. Thank you for your time and expertise.

Just ordered three 16TB drives from OWC with a new Mercury Elite Pro Dual enclosure (plus some older singles I have around the house). I'm on the end of a 4-year build of a website with massive amounts of content and needed 3 more disks for backup. Since I'm hoping to continue still working while certifying the new disks, I thought "I'll just use my Windows machine to certify with the other enclosures and keep my Mac free to work on with the ThunderBay attached." 

But when I installed SoftRaid on Windows 10, I can Certify a disk alright, but it gives me no control or option to select how many passes I want to go over the disk during certification. It looks like the default is "2." And there is nowhere with the other options I'm used to in the Mac version. 

Is the Windows version just wimpier or is there some options menu I haven't found yet? I was planning on buying a second copy when the Windows trial is up, but if it's a wimpier version, I need to know to restructure. Thanks,

Sean 

I've purchased Flex 8 last year and set it up as a single RAID 5 volume with 8x8TB HGST drives and it's been working fine since. However, whenever I try to Validate a RAID 5 volume, the speed is pretty extreme like shown in the screenshot (and the ETA gets longer as it progresses)...

I've tried many times with different time of the day (in case my system is using the RAID during certain time), and even have re-set a brand new volume just yesterday, with only 1TB of data in it.

Wondering if I'm missing anything here? Here are the details of my setup:

RAID 5 - 8x8TB HGST performance drives

Optimized for Digital Video

64k

System Specs:

Tried both Windows 10 Pro / 11 Pro

i7-12900KF

Asus TUF H670-Pro WiFi D4 with

ThunderboltEX4 add-on card

128GB DDR4 RAM

GeForce RTX 3090

Thank you.

I'm using an older version of the SoftRAID software (version 3.0.8.8) because newer versions won't let certain applications write to the RAID.  (See thread at https://forums.softraid.com/windows10-issues/major-incompatibility-with-numerous-applications/)

Today, Windows 11 gave me a warning that I should restart my computer to allow some issues with disk data to be resolved.  Upon restarting, I opened up SoftRAID to make sure there was nothing wrong, but lo-and-behold I found two errors on my RAID...  (1) Your Data Isn't Protected since there's a disk missing and (2) Data is out-of-sync due to changes made to the RAID while a member was missing.  The RAID's member disks show that Disk #2 (slot B in my OWC Thunderbay 8) is missing...  However, under "All Disks" the disk is there and active and seems to be fine, though it isn't reported as being part of the RAID.

I'm hesitant to "rebuild" the RAID as that window suggests, particularly if it doesn't see that "missing" disk...  but not sure how to get that disk seen.  Several of the disks haven't wanted to show any disk health info since Day 1 -- so I'm at least running validation on that "missing disk" right now, though it says it'll take 18+ hours at this point.  With over 100 TB on this RAID, I want to make sure I do what I can to protect and resolve any issue ASAP.  Please help! 

(As for back ups -- I am using Backblaze, but it's only half backed-up -- the most important parts for now -- and downloading that much replacement data would be a far distant option from simply fixing the RAID).

Is this the expected behavior?  Why does the speed change so much? Just as a test making the disk raid 0 I am seeing 900mb per second writes.