Running mission-critical Microsoft SQL Server databases on Amazon FSx for NetApp ONTAP

This article provides comprehensive best practices for running mission-critical SQL Server databases on Amazon FSx for NetApp ONTAP, covering storage configuration, performance optimization, and cost efficiency.

• FSx for ONTAP combines AWS infrastructure with NetApp storage, delivering sub-millisecond latencies and millions of IOPS per file system
• Performance Tier SSD hosts active SQL Server files; Capacity Pool stores cold/archive data cost-effectively with seamless tiering
• Size Performance Tier at 1.35x total database size to ensure headroom for snapshots and enterprise features
• Storage Virtual Machines (SVMs) provide multi-tenant isolation with separate credentials, performance domains, and network endpoints
• Maintain 1:1 volume-to-LUN ratio for granular backup/restore and performance isolation across databases
• Separate data files (.mdf/.ndf), transaction logs (.ldf), and TempDB onto dedicated volumes to prevent I/O contention
• Five storage layout scenarios: single database, database consolidation, filegroups, read-only filegroups with tiering, and TempDB placement
• Tiering policies (snapshot-only, auto) move inactive filegroups to Capacity Pool, reducing costs while keeping active data on SSD
• Stack multiple FSx for ONTAP file systems for workloads exceeding single-system performance limits

Proper database layout, volume-to-LUN ratios, and tiering strategies create a scalable, resilient, and cost-effective foundation for mission-critical SQL Server workloads on AWS.