Windows Recovery Fix: KB5066189 OOB Update (Aug 2025)

Microsoft has acknowledged a critical reliability problem introduced by its August 12, 2025 Windows security rollup that disabled core recovery features on multiple client builds of Windows 10 and Windows 11, and issued an out‑of‑band (OOB) cumulative update — KB5066189 — to resolve failed Reset my PC, Fix problems using Windows Update, and RemoteWipe operations for affected systems. (support.microsoft.com)

Background​

Microsoft’s August security updates (released August 12, 2025) were distributed as the monthly cumulative rollups for consumer and enterprise channels. For Windows 11, the relevant packages included KB5063875 for Windows 11 versions 23H2 and 22H2 (OS Builds 22621.5768 and 22631.5768) and KB5063878 for Windows 11 24H2 (OS Build 26100.4946). These packages carried the usual mix of security fixes and quality improvements, but reports surfaced quickly that some of the August patches were introducing new, highly visible problems on certain systems. (support.microsoft.com)
Within days, users and enterprise administrators began reporting two distinct but serious issues: (1) recovery and reset routines failing on systems updated with the August rollups, and (2) reports of storage drives—particularly some SSDs—disappearing after heavy write operations following installation of the 24H2 patch. Microsoft has publicly acknowledged the recovery/reset failures and released a targeted OOB update to correct them. The storage disappearance reports are being investigated and have produced mixed telemetry and vendor responses to date. (bleepingcomputer.com)

---

What Microsoft confirmed (the hard facts)​

• Microsoft confirmed that after installing the August 2025 security update on affected client platforms, attempts to reset or recover the device might fail — specifically impacting the Reset my PC flow, the Fix problems using Windows Update recovery flow, and the RemoteWipe CSP used for remote resets. Microsoft documented this in its Windows release health and KB updates and flagged the issue as a known problem introduced by the August rollups. (support.microsoft.com, bleepingcomputer.com)
• To remediate the failure, Microsoft published a non‑security, out‑of‑band cumulative update KB5066189 on August 18–19, 2025 depending on the release channel, which raises the affected Windows 11 builds to 22621.5771 / 22631.5771 and includes a specific fix for the reset/recovery failure. This OOB update supersedes the problematic components of the August security update for the impacted platforms. Microsoft recommended deploying KB5066189 instead of the original August LCU if systems had not yet been updated. (support.microsoft.com, bleepingcomputer.com)
• The known reset/recovery problem affected client platforms that had received their respective August LCUs, including Windows 11 versions 23H2 and 22H2 (via KB5063875), and several Windows 10 flavors covered by related KBs. Microsoft’s guidance was explicit: if your organization uses the impacted platforms and has not yet deployed the August patch, apply the OOB update instead. (bleepingcomputer.com, support.microsoft.com)

These are Microsoft’s official positions and the basis for IT teams’ immediate response plans. (support.microsoft.com)

---

Timeline — concise sequence of events​

• August 12, 2025 — Microsoft releases the August monthly quality/security updates (including KB5063875 and KB5063878) across Windows channels. (support.microsoft.com)
• Mid‑August 2025 — Users report Reset my PC / recovery failures and also a separate set of reports where SSDs/HDDs disappear during heavy writes on some 24H2 systems. Independent outlets and forums begin aggregating reports. (itpro.com, windowscentral.com)
• August 18–19, 2025 — Microsoft posts a Windows release health notice acknowledging the reset/recovery problem, then issues the out‑of‑band fix KB5066189 (and companion OOBs for other affected Windows 10 branches). Microsoft advises installing the OOB update where appropriate. (support.microsoft.com, bleepingcomputer.com)

---

Technical analysis: what likely went wrong​

The recovery/reset flows in Windows are layered processes that interact with the OS image servicing stack, the Windows Recovery Environment (WinRE), platform drivers, and the Storage & Boot stacks. A flaw in any component — especially one that touches the servicing stack, WindowsRE mount/unmount operations, or the remote reset APIs — can cause reset attempts to fail or repeatedly roll back.

• The reset and recovery routines rely on a stable servicing stack and consistent OS image payloads to perform file replacements, registry cleanup, and user profile preservation when requested. When a recent cumulative update changes a servicing component in a way that WinRE or the reset APIs do not expect, the operation can detect mismatched state information during reinstall/rollback and abort. Microsoft’s OOB update explicitly targets servicing/LCU components, which aligns with this failure mode. (support.microsoft.com)
• The fact that Microsoft released a cumulative OOB (rather than a driver-only or WinRE‑only patch) suggests the root cause was inside the combined servicing/LCU layer rather than a third‑party driver. OOB cumulative updates replace problematic LCU payloads and shift the system to a corrected build number (22621.5771), which is consistent with resolving mismatched servicing state. (support.microsoft.com)
• The separate storage disappearances reported on 24H2 (KB5063878) appear to be tied to heavy continuous write workloads on drives already at high utilization and may implicate SSD controller firmware interactions, OS write buffering, or metadata/cache handling under extreme conditions. Reports point toward some Phison‑controller SSDs being disproportionately represented in user threads and tests, but Microsoft has not publicly asserted a definitive universal cause and stated it is investigating with partners. This distinction—between a confirmed reset/recovery bug and an ongoing storage investigation—is critical. (bleepingcomputer.com, windowscentral.com)

Caution: unless Microsoft or a controller vendor publishes a root cause report that includes a reproduction case and telemetry, the storage failures remain under investigation and should be treated as an active risk rather than a closed issue. (bleepingcomputer.com)

---

Who and what is affected​

• Impacted recovery/reset flows: Reset this PC, Fix problems using Windows Update in System Recovery, and RemoteWipe CSP. These are used by consumers and IT admins alike when reimaging or remotely wiping devices. (support.microsoft.com)
• Affected platforms for the reset/recovery bug include Windows 11 23H2 and 22H2 (KB5063875), and corresponding Windows 10 builds that received related August updates. Microsoft’s KB and release health guidance call out these client SKUs specifically. Windows 11 24H2 was reported as unaffected for the reset bug but has separate storage reports tied to KB5063878. (bleepingcomputer.com, support.microsoft.com)
• Scope: Microsoft’s initial advisory covers client OS versions — server SKUs were not highlighted in the reset advisory — and telemetry so far has not indicated broad server impact. Administrators managing fleets should prioritize endpoint and client remediation planning first. (support.microsoft.com)

---

Independent confirmation and broader reporting​

Multiple independent outlets and research communities documented both the reset failure and the storage disappearance reports. Investigative and reporting outlets confirmed Microsoft’s acknowledgement and tracked the OOB release timeline, while storage anomalies were escalated by power users and SSD researchers. The important corroborations:

• News outlets and technical media reported Microsoft’s confirmation of broken reset/recovery features and the release of KB5066189 as the fix. These outlets reiterated Microsoft’s recommendation to apply the OOB update instead of the August LCU for affected systems. (techradar.com, bleepingcomputer.com)
• Independent storage reporting and hands‑on tests suggested the storage disappearance events often occur during heavy writes (roughly 50 GB or more) to drives with high utilization (often >60%) and highlighted Phison‑controller drives as frequently appearing in reports. Microsoft has engaged vendors and requested telemetry from affected customers but had not initially reproduced a systemic failure in its internal testing at the time of public updates. (windowscentral.com, bleepingcomputer.com)

This independent confirmation underscores two practical realities: the reset/recovery regression is a confirmed, actionable bug that Microsoft patched with an OOB update; the SSD/hard‑drive disappearance story is real in a set of user reports but remained under investigation and not universally reproducible at the time Microsoft published its advisories. (support.microsoft.com, windowscentral.com)

---

Practical recommendations — immediate actions for users and admins​

For home users, IT administrators, and managed service teams, the following steps are prioritized to minimize disruption and data risk.

• Install the OOB update if you have already applied the August security update and are on an affected client SKU. Microsoft’s OOB (KB5066189) supersedes the problematic August LCU for the affected builds and requires a restart. If the system is impacted by the reset/recovery failure, this is the recommended path. (support.microsoft.com)
• If you have not yet installed the August 2025 monthly security updates and your systems fall into the affected SKUs list, apply KB5066189 (or the equivalent OOB for your Windows 10 branch) instead of KB5063875. Microsoft explicitly suggested organizations do this to avoid the issue. (bleepingcomputer.com, support.microsoft.com)
• For systems on Windows 11 24H2 (KB5063878) or for any machine with an SSD that is heavily utilized, avoid large continuous write operations (e.g., large game updates, mass file moves, backup‑to‑same‑disk) until the storage investigation produces clearer guidance or fixes are released. Back up critical data immediately using external media or cloud backups. (windowscentral.com, bleepingcomputer.com)
• For managed fleets, test the OOB update in a staging ring before broad deployment. Maintain a rollback plan and ensure recovery images are available offline (bootable rescue media) in case a device cannot be reset remotely before remediation is deployed. Logging and telemetry collection from affected endpoints will be important for post‑mortem analysis. (support.microsoft.com)
• If you experience storage disappearance or corruption after applying an update, collect system logs, event viewer entries, SMART reports, and file operation traces where possible, and file a Feedback Hub report or contact vendor support. Microsoft has requested affected users provide details for its investigation. (bleepingcomputer.com)

---

Step‑by‑step: how to get KB5066189 (concise guide)​

• Open Settings -> Windows Update and choose Check for updates. If KB5066189 is available for your device, it should appear as an optional (out‑of‑band) update. Install and restart when prompted. (support.microsoft.com)
• For enterprises using WSUS or SCCM/ConfigMgr, approve the KB5066189 package for the applicable device groups; ensure your servicing stack is up to date so the combined package can install cleanly. The OOB is cumulative and does not require previous LCUs beyond the servicing prerequisites already documented. (support.microsoft.com)
• If you require manual download, obtain the update from the Microsoft Update Catalog for the precise OS build and architecture, and install using the standalone MSU file with a reboot. Microsoft’s KB pages include file information for offline installs. (support.microsoft.com)

Note: the OOB update contains no additional security fixes beyond the August LCU; it is a non‑security quality update intended to correct functionality. (support.microsoft.com)

---

Risk assessment and enterprise impact​

• Operational risk: The reset/recovery issue affects fundamental recovery paths that enterprises and help desks rely on for device reprovisioning and remediation. This increases help desk workload, especially for remote devices where RemoteWipe/remote reset is used to remediate compromised or non‑responsive endpoints. The OOB update mitigates this but organizations that applied the August LCU broadly should verify device readiness and consider a controlled remediation rollout. (support.microsoft.com)
• Data risk: The immediate reset bug does not, in Microsoft’s advisory, intentionally corrupt user data — it prevents the reset operation from completing properly. The separate storage disappearance reports, however, raise a possible data loss risk for a subset of devices during heavy writes; this represents a genuine data integrity concern and justifies aggressive backup discipline until vendors confirm root cause and remediation. (bleepingcomputer.com, windowscentral.com)
• Reputational and supply risk: If storage controller vendors and OS vendors must coordinate firmware and OS fixes, remediation timelines can vary by vendor, model, and firmware versions. Enterprises with diverse hardware fleets should catalog vulnerable models and track vendor firmware updates closely. (windowscentral.com)

---

Why these bugs slipped through — process and testing gap analysis​

• Regression testing often covers typical workflows; edge cases that trigger recovery or exceptional storage behavior may be exercised less frequently in standard telemetry or automated test suites. Reset/recovery and rare heavy‑write controller interactions can therefore avoid detection until broad real‑world installs surface them. The speed of modern cumulative servicing and the huge variety of client hardware combinations create testing blind spots. (support.microsoft.com, windowscentral.com)
• The complexity of the Windows servicing stack — combining SSU, LCU, optional features, OEM drivers, and firmware interactions — means a change that looks trivial in isolation can produce emergent failures when deployed across millions of hardware permutations. This incident highlights the need for expanded scenario testing for recovery flows and increased collaboration with storage partners on stress‑test harnesses. (support.microsoft.com, bleepingcomputer.com)

---

What to monitor next​

• Watch for subsequent Microsoft advisories or updated KB articles that expand affected SKUs, publish root cause analyses, or release additional OOB or security updates. Microsoft has a habit of publishing follow‑up notes in Windows release health once telemetry and vendor investigations conclude. (support.microsoft.com)
• Monitor SSD vendor firmware bulletins and advisories. If storage vendors publish firmware updates or guidance tied to the August patches, apply vendor‑recommended firmware updates per your change control processes. (bleepingcomputer.com)
• Review enterprise telemetry for failed recovery attempts, SRT/WIM mount errors, and recent mass‑write jobs to detect early indicators of the storage anomaly. Logging these events proactively will speed incident response if more devices become affected. (windowscentral.com)

---

Final assessment — strengths, limitations, and risks​

Strengths:

• Microsoft’s quick acknowledgement and the rapid issuance of an out‑of‑band cumulative update show a responsive servicing pipeline; for the confirmed reset/recovery regression, the OOB patch is an effective corrective action and is straightforward to deploy. (support.microsoft.com)
• Transparency about build rollups, corrected build numbers (22621.5771 / 22631.5771), and explicit guidance to install the OOB rather than the problematic LCU helps IT teams make deterministic remediation decisions. (support.microsoft.com)

Limitations and risks:

• The storage disappearance issue remains less well defined: it is backed by multiple community reports and vendor engagement but lacked a single, authoritative root‑cause statement at the time Microsoft published the OOB fix for the reset bug. This ongoing investigation poses continuing data integrity risk for certain drive models and workloads. (bleepingcomputer.com, windowscentral.com)
• The incident underscores the broader challenge of delivering monthly cumulative patches across an enormous combinatorial space of PC hardware and firmware. Even with rapid OOB patches, enterprises must treat major monthly rollouts as potential operational hazards and keep conservative staging and backup practices. (support.microsoft.com, windowscentral.com)

Cautionary note: any specific claims about particular SSD models or definitive blame on any controller or firmware should be treated as provisional until vendors and Microsoft publish coordinated technical post‑mortems; many community reports are anecdotal or from narrow testbeds and may not represent global telemetry. (pcworld.com, amagicsoft.com)

---

Conclusion​

The August 2025 Windows monthly updates produced a confirmed regression that broke key recovery pathways for affected Windows client builds; Microsoft acknowledged the problem and issued an out‑of‑band cumulative update (KB5066189) to repair Reset and recovery functionality. That OOB update is the recommended fix for impacted systems and should be used instead of the original August cumulative if systems have not yet been updated. (support.microsoft.com, bleepingcomputer.com)
At the same time, a separate and more troubling set of reports about SSDs and HDDs disappearing under heavy write conditions after the 24H2 cumulative (KB5063878) remains under active investigation, with vendor and Microsoft involvement but without a definitive universal cause at the time of reporting. Until that investigation concludes, administrators should prioritize backups, avoid heavy write operations on at‑risk systems, and track firmware advisories from storage vendors. (windowscentral.com, bleepingcomputer.com)
For IT teams, the immediate win is clear: install KB5066189 (or the appropriate OOB for your branch) for affected clients, validate recovery flows in a test ring, and treat broad monthly rollouts with staged deployments and rigorous preupdate backups. For home users, install the OOB if you need reliable reset/recovery paths, and back up any important data — especially if your system uses an SSD model that has appeared in the storage reports. (support.microsoft.com, windowscentral.com)
The incident is a reminder that even routine security updates can have systemic side effects in a complex hardware ecosystem; disciplined deployment, layered backups, and close vendor coordination remain essential defenses against update‑related disruptions. (support.microsoft.com, bleepingcomputer.com)

---

Source: WinCentral Microsoft confirms 1 Windows 11 23H2 issues post KB5063875