Windows Server September 2026 RDS Patch Break: What UK Admins Should Do Now

What's actually breaking on RDS servers

Reports gathered by BleepingComputer describe a consistent pattern: Remote Desktop Services work normally for a few hours after the September 2026 cumulative update installs, then connections start failing. Existing sessions can get stuck unable to log off or disconnect, while new connection attempts hang during negotiation before timing out entirely. Several administrators say a standard reboot does not clear the fault, and a hard reset of the server is the only way to restore service.

One reader told BleepingComputer that in their environment "all terminal servers... are failing (sessions dropping, no new connections possible, only solution is a hard reset) within a day." Another said the problem only surfaces after the first user logs out, at which point services crash and no further sign-ins succeed — even after a restart. The specific culprits named are KB5122876 on Windows Server 2019, KB5122882 on Windows Server 2022, and KB5122871 on Windows Server 2025, meaning this is not confined to a single OS branch.

A plausible root cause, but no official confirmation yet

One Server 2022 administrator investigating the fault reported what looked like a deadlock between the Remote Desktop service and the Local Session Manager, pointing to a hang in a specific RDP server routine with no timeout configured. Microsoft has not confirmed this as the underlying cause. The company has told BleepingComputer only that it is aware of the reports, is investigating, and will issue guidance once it is available — leaving affected estates without an official fix timeline for now.

Rollback restores RDS, but strips out this month's fixes

Admins who have removed the September update say Remote Desktop functionality returns to normal. The catch is significant: rolling back also removes every security fix shipped in this month's release, including fixes for two actively exploited zero-days among the record 966 flaws closed by Microsoft's September 2026 Patch Tuesday. Any organisation reverting the update on RDS-facing servers is knowingly reopening that entire patch set, including live exploitation paths, purely to keep remote sessions functional.

This lands in the same cycle where Microsoft separately warned that Windows Server 2025 memory-management changes can trigger application crashes — a reminder that this month's Server-side regressions are not isolated to one component. For estates already juggling multiple post-patch issues, that raises the operational cost of any single rollback decision.

Practical steps while Microsoft investigates

Until Microsoft confirms a fix, treat this as a live production risk rather than a wait-and-see patch note. UK admins should validate exposure on every RDS-facing host running the affected KBs before users start reporting outages mid-shift.

• •Test the September update in a non-production RDS pool first and monitor session behaviour for several hours, not minutes, before wider rollout
• •Where RDS is business-critical, weigh a temporary rollback against the newly reopened CVEs, including the two zero-days, on a case-by-case basis
• •Document a hard-reset procedure and escalation path for on-call teams, since a normal reboot is not reliably clearing the fault
• •Keep exploited-vulnerability patching current on non-RDS systems even if RDS hosts are held back, to avoid blanket exposure
• •If in-house capacity is stretched, third-party maintenance options can cover emergency response while Microsoft finalises guidance

Why this matters beyond one patch cycle

This is at least the latest in a run of Remote Desktop regressions tied to Windows updates, and Microsoft has previously used Known Issue Rollback to resolve similar RDP faults on managed devices — suggesting a targeted fix is a realistic outcome rather than a full re-engineering effort. But the recurrence itself is the signal UK buyers should note: RDS and Terminal Server estates keep surfacing as a fragile point in Microsoft's update pipeline, and Windows Server 2022 also reaches end of mainstream support within 60 days, adding pressure to plan the next platform move rather than firefight this one indefinitely.

Every hour of RDS downtime has a direct cost in lost productivity and support overtime, so it is worth being able to calculate the cost of downtime before deciding whether to roll back, hold the patch, or accept a short outage window. Organisations already migrating to Windows Server 2025 should factor this regression into cutover testing rather than assuming a clean baseline. For teams reassessing wider resilience, our cyber security services, including managed detection & response and zero trust approaches, can reduce reliance on any single remote-access path during incidents like this.