After Fixing WD BSODs, Windows 11 24H2 Update Makes SSDs Disappear and Corrupt Data

Western Digital has finally released firmware fixes for the dreaded Blue Screen of Death (BSOD) loops that plagued its NVMe SSDs after installing Windows 11 24H2. But as that drama subsides, a new August 12, 2025 cumulative update—KB5063878—is causing a different nightmare: drives can catastrophically vanish during heavy write workloads, taking in-flight data with them.

Two distinct storage crises now trail Microsoft’s flagship OS update, exposing fragile interactions between the Windows storage stack and SSD controller firmware. For users and administrators alike, the message is stark: backup aggressively, update firmware before applying OS patches, and watch for signs of drive instability.

A Tale of Two Storage Regression Bugs

Windows 11 version 24H2, the major 2024 feature update, introduced changes at multiple layers of the storage stack. For DRAM-less NVMe SSDs that rely on Host Memory Buffer (HMB) to borrow system RAM for caching, those changes proved devastating. Two failure patterns emerged months apart, each tied to subtle host-side behavior shifts.

The October 2024 HMB BSOD Crisis

Shortly after 24H2’s broad rollout, owners of specific Western Digital and SanDisk NVMe models reported relentless BSOD loops. The most common victims were the WD Black SN770, WD Blue SN580, and SanDisk Extreme M.2, especially in 2TB capacities. Error logs pointed to “Critical Process Has Died” and storage controller faults.

Community sleuths quickly identified the trigger: a change in HMB allocation size. Drives that previously received around 64 MB of HMB were now being assigned up to 200 MB—a window that older firmware couldn’t handle, crashing the controller and taking Windows down with it. A registry workaround that limited or disabled HMB via HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\StorPort\HmbAllocationPolicy became a stopgap, but it reduced SSD performance and carried the usual registry risks.

Western Digital and SanDisk responded with targeted firmware updates. On affected models, users were strongly urged to update immediately through the Western Digital Dashboard or SanDisk Dashboard tools. Microsoft also blocked upgrades on systems with out-of-date firmware, preventing further automatic distribution of 24H2. The table below lists the confirmed models and required firmware versions:

Users were warned that firmware updates could cause data loss, a grim reminder to back up critical files first. For most, however, the firmware resolved the BSOD loops completely.

August 2025: KB5063878 Triggers Drive Disappearances

On August 12, 2025, Microsoft released cumulative update KB5063878 (OS Build 26100.4946) for Windows 11 24H2. The official KB page initially listed no known issues. Then independent testers and outlets like Guru3D and Tom’s Hardware began reproducing a disturbing new regression.

Under sustained, large sequential writes—community repros often used transfer volumes of 50 GB or more—certain NVMe SSDs stop responding entirely. The drives vanish from Device Manager and Disk Management, become unreadable to SMART or vendor diagnostic tools, and in a subset of cases, files written during the failure window are corrupted or go missing. A reboot may restore drive visibility but does not guarantee the integrity of the interrupted writes.

Affected hardware extends beyond a single brand. Many failing drives use Phison controller families or DRAM-less designs, though community lists are not exhaustive. The working theory points to a host-side timing or buffering regression in the cumulative update that causes the SSD controller to stall or crash under heavy I/O. Unlike the earlier BSODs, which surfaced immediately after the OS upgrade, this failure is workload-dependent and can ambush users during routine large file transfers.

Why Two Different Failures from the Same Update?

Both incidents share a common root: Windows 24H2’s storage stack changes expose latent firmware bugs in SSD controllers. HMB negotiation, command timing, buffer sizes, and DMA behavior all interact tightly between the OS (stornvme, StorPort) and the drive’s firmware. A minor shift can push a controller into an edge condition it wasn’t designed to handle.

In the October BSOD case, the larger HMB allocation caused the controller to crash immediately. In the August scenario, sustained writes stress the SSD’s internal metadata and mapping operations; a subtle host-side regression then causes the controller to become unresponsive, leading to a device drop-out. The result is not just an annoyance—it’s a data integrity emergency.

Microsoft and Vendor Response: Mixed Messaging

Western Digital and SanDisk earned praise for delivering firmware fixes and dashboard tools. Their advisories explicitly warn users to back up data before applying firmware, and Microsoft’s upgrade blocks prevented many at-risk systems from receiving 24H2 in the first place.

For KB5063878, however, the communication gap has been glaring. While community reproduction and enthusiast press highlighted the vanishing-drive issue within days, Microsoft’s KB page continued to list no known issues, leaving users unaware of the danger. Some enterprise deployments also hit installation regressions (error 0x80240069), adding to admin frustration. The discrepancy between published KB information and real-world experience has drawn sharp criticism.

Practical Guidance: What Users and Admins Must Do Now

Home User Checklist

1. Back up critical data now. Use an external drive or cloud service before updating firmware, installing updates, or performing large transfers.
2. Identify your SSD model. Check Device Manager or use vendor tools (WD Dashboard, SanDisk Dashboard) to see model and firmware version.
3. Update SSD firmware first. If your drive is on the vendor’s list (see table above), apply the firmware update immediately. Follow vendor instructions precisely and expect a mandatory shutdown to complete the process.
4. If you’ve already installed KB5063878 and rely on heavy writes, postpone large file transfers until vendors confirm compatibility. Watch for official fixes or Known Issue Rollback (KIR) mitigations.
5. Registry workaround (last resort). If you face BSODs and a firmware update isn’t available, edit HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\StorPort\HmbAllocationPolicy to limit or disable HMB. Note that this reduces performance and should only be used temporarily. Always backup the registry first.

If a Drive Disappears Mid-Transfer

• Stop all writes immediately. Do not retry the transfer.
• Check Device Manager and Disk Management. If missing, perform a graceful reboot—but realize that file integrity may already be compromised.
• Use vendor tools to query SMART and firmware. If telemetry is unreadable, suspect corruption.
• Restore from backup if files are lost. If you lack backups, image the disk sector-by-sector before any further writes, then consult professional recovery services.

Enterprise Administrator Flags

• Stage and test KB5063878 in isolated environments before broad deployment. Validate with typical workload profiles—backups, large deployments, imaging, AV scans.
• Monitor WSUS/SCCM for installation errors (especially 0x80240069) and storage-related regressions.
• Have a rollback plan that includes restoring affected files from backup, not just removing the patch.

Strengths and Weaknesses in the Response

What went right: Vendor responsiveness was swift for the BSOD issue. Western Digital and SanDisk produced effective firmware, and the upgrade blocks worked. Community engineering played a vital role, accelerating awareness and mitigations.

What went poorly: The KB5063878 communication gap risked data loss for unaware users. The initial “no known issues” stance on Microsoft’s KB page clashed with verifiable community evidence, and the reliance on Insider and staged rollouts didn’t catch the heavy-write regression. The failure mode—potential corruption and unrecoverable files—raises the stakes far beyond a minor stability blip.

Risk Assessment: How Worried Should You Be?

If you own a modern NVMe drive from a mainstream brand, the probability of hitting either issue depends on your model, firmware, and workload. The initial BSODs concentrated on specific WD/SanDisk 2TB models—those are now largely resolved with firmware. The KB5063878 regression affects a broader set of drives, especially those with Phison controllers, but remains non-universal.

For affected users, the impact is severe: repeated crashes, vanished drives, and potential file corruption. The conservative approach for anyone handling critical data is to back up now, avoid large writes on systems with the August update, and wait for clear compatibility confirmations.

Conclusion

The Windows 11 24H2 storage saga is a sobering reminder that OS patches—even routine security updates—can interact fatally with hardware. Two separate regression bugs, months apart, both trace back to changes in how Windows talks to SSDs. Firmware fixes have tamed the BSODs for Western Digital drives, but the newer vanishing-drive bug underscores ongoing risk.

Prioritize backups, keep firmware current, and stage major updates carefully. This double-whammy should also push Microsoft to improve its real-world testing and communication—because when storage falters, the data you’ve trusted Windows to protect can disappear in an instant.