As reported by BleepingComputer, Microsoft has confirmed that Windows updates released since the October 2026 Patch Tuesday are causing SQL Server performance and stability degradation due to a connection-leak bug in linked server queries using OLE DB providers. While this is not a security vulnerability, the operational impact on data-dependent environments warrants immediate attention from database and infrastructure teams.
The Real Risk: Slow-Burn Degradation, Not a Sudden Crash
What makes this issue particularly insidious is its cumulative nature. Connections to external databases — including Oracle and PostgreSQL linked servers — remain open after queries complete and accumulate over time rather than being released. This is a resource exhaustion pattern that may not trigger traditional alerting thresholds until performance has already noticeably degraded. Unlike a crash or service failure that generates immediate, obvious alerts, this bug manifests as a gradual decline: slower query responses, increasing memory pressure, and eventual instability.
For environments running automated ETL pipelines, reporting workloads, or applications with high-frequency cross-database queries, the connection accumulation rate could be rapid enough to cause meaningful degradation within hours rather than days. Smaller deployments with infrequent linked server usage may not notice symptoms for longer periods, creating a false sense of security.
This is a textbook example of why patch management programs must include post-deployment health monitoring — not just validation that patches installed successfully, but verification that dependent systems continue performing within expected parameters.
Who Is Most Exposed
Operational Context: A Pattern of Quality Issues
This is not an isolated incident. The article notes a separate SQL Server crash issue on Windows Server 2025 related to the Lock Pages in Memory (LPIM) policy, and deployment failures with the November 2026 security updates on systems where the October preview was installed. Taken together, these issues suggest a quality control gap in Microsoft's recent update pipeline that should give IT teams pause before aggressive patch rollout — particularly for critical database infrastructure.
Shield53 Recommendations
- Audit linked server usage immediately: Query
sys.serversand monitorsys.dm_exec_connectionsfor external database connections that remain open after associated queries complete. Establish a baseline now if you don't have one. - Implement proactive monitoring: Set alerts for connection count growth trends on linked server targets, not just absolute thresholds. A steady upward trend with no corresponding active query load is the signature of this bug.
- Request the Microsoft Support mitigation: Microsoft has indicated a private mitigation is available through Support for Business. Open a support case now if linked server workloads are critical to your operations — don't wait for the out-of-band update.
- Plan the OOB update deployment: When the out-of-band fix arrives, prioritize database servers in your deployment sequence. Test on non-production SQL instances first, specifically validating linked server connection lifecycle behavior.
- Document the workaround limitations: Manual connection closure and SQL service restarts are temporary measures that may kill active queries. Coordinate with application teams before executing, and schedule restarts during maintenance windows where possible.
- Review your patch-tiering strategy: If your organization applied October Patch Tuesday updates to database servers without a staging period, reconsider your rollout cadence. Critical database infrastructure benefits from a 1-2 week staged deployment with health validation between tiers.
Microsoft has stated the out-of-band fix will arrive within days. Until then, database administrators should treat this as an active operational risk and monitor accordingly — the worst outcome is discovering the problem only after end users report it.
