SQL-Cluster-Sentry

The "Why"

This repository was born from a specific operational failure. In a previous role, a lack of distinction between Local High Availability (HA) and Geographic Disaster Recovery (DR) led to an accidental failover of a 4.5TB Data Warehouse across a 1GB WAN link.

The root cause was treating a Failover Cluster Instance (FCI) like an Availability Group (AG).

The Solution

SQL-Cluster-Sentry is a PowerShell-based logic gate designed to be run before any automated patching or maintenance cycle. It acts as a "Traffic Cop" for the cluster.

Core Logic

The script performs a Topology Audit before allowing a move:

1. Who am I? Determines if the instance is an FCI (Shared Storage) or AG (Replicated Storage).
2. Where is my partner? Compares the Active Node's Subnet/Site against the Target Node.
3. The "Kill Switch": * IF InstanceType == FCI
   • AND SourceSubnet != TargetSubnet
   • THEN BLOCK FAILOVER.

Intended Workflow

1. MECM/SCCM triggers the maintenance window.
2. Test-SqlFailoverSafety.ps1 runs.
3. If returns TRUE: Proceed with standard Cluster-Aware Update.
4. If returns FALSE: Abort patch on this node. Alert SysAdmin. Do Not Move Data.

Philosophy

"Automation without Context is just automated destruction." - Gavin Dobbs