Use the primary SSD tier for FSx for ONTAP volumes for optimal performance. The capacity pool tier might reduce performance and increase latency. The primary SSD tier is designed for high-performance workloads, while the capacity pool tier is optimized for cost efficiency and might not meet the performance requirements of SQL Server workloads.

Maintain enough file system capacity above your total volume size. Too little free space causes performance issues; too much wastes storage and increases costs.

Size and monitor your SQL Server log drive to prevent transaction rollbacks, database unavailability, data corruption, and performance issues caused by a full log drive.

Size and monitor TempDB to maintain performance and stability. Too little TempDB space can cause query slowdowns, application timeouts, and system crashes.

Separate data and log files onto different drives to improve parallel I/O, backup flexibility, and restore behavior. We recommend separating data and log LUN paths into different volumes for smaller databases.

Place TempDB on a dedicated drive to avoid I/O contention and improve SQL Server performance and stability.

Configure thin provisioning for FSx for ONTAP volumes to make better use of storage and reduce costs. Thin provisioning lets you allocate storage as needed instead of reserving all capacity at once.

Turn on autosize and set the mode to grow so volumes increase automatically before they become full and read-only.

Configure block device space settings for LUNs used by Microsoft SQL server instances to prevent write failures and improve space efficiency on FSx for ONTAP.

Set snapshot copy reserve to 0% and enable snapshot autodelete to protect space for active data.

When space runs low, expand the volume before deleting snapshot copies. Expanding the volume prevents "volume full" errors, reduces write failures, and preserves snapshot recovery points.

Use the snapshot-only tiering policy, which moves only snapshot data to the capacity tier while keeping active data on the SSD tier. This approach preserves low-latency performance for SQL workloads while reducing storage costs. Tier snapshot data after a 7-day cooling period.

Enable deduplication, compression, and compaction to reduce storage use and cost while maintaining performance.

Match the ONTAP LUN Operating System value to the host partitioning scheme to maintain correct I/O alignment.

Configure Multipath I/O (MPIO) for resiliency and throughput. Use five MPIO sessions per target interface for SQL Server on iSCSI LUNs, monitor path status, and set host timeout to 60 seconds.

Set NTFS allocation unit size to 64K to improve disk efficiency and reduce fragmentation for SQL Server workloads.

Disable scheduled snapshots for FSx for ONTAP volumes used by Microsoft SQL Server. Instead, manage snapshots externally using tools such as NetApp SnapCenter, which create application-consistent backups and help prevent data corruption during restore operations.

Rightsize your SQL Server EC2 instance based on workload demand. Increase size when under-provisioned and reduce size when over-provisioned to balance cost and performance.

Apply the latest operating system patches to improve security and reliability.

Configure Receive Side Scaling (RSS) to distribute network processing across CPUs to improve network performance. Validate settings before production rollout.

Align EC2 Maximum Transmission Unit (MTU) settings with FSx for ONTAP paths to prevent fragmentation and maintain stable throughput.

Review SQL Server license usage at the host level. If paid features are not in use, the license is not optimized and may increase cost.

Apply the latest SQL Server patches to improve security and system reliability.

Configure MAXDOP to optimize query performance. Values of 4, 8, or 16 typically work well. Test your workload and monitor for parallelism wait types such as CXPACKET.

Use application-consistent snapshots with NetApp SnapCenter to take accurate, reliable snapshots of your volume data at a specific moment in time. SnapCenter makes backups easier and helps you restore data quickly, reducing downtime and protecting critical workloads.

Enable cross-region replication (CRR) to improve data availability and disaster recovery in your SQL Server environment. CRR helps protect against regional outages and ensures business continuity by replicating data across different locations.

Configure FSx for ONTAP backups or AWS Backup to meet retention and compliance requirements while avoiding duplicate backup workflows.

Configure quorum correctly for SQL Server Failover Cluster Instances so the cluster can tolerate expected node failures.

Use the same drive letters on every node in your SQL Server high availability setup. This makes management easier and helps failover work correctly.

Tune heartbeat settings to avoid unnecessary failovers while still detecting real failures quickly.

All shared disks (iSCSI LUNs) must be accessible from both nodes in the FCI deployment model so failover can work.

Validate service account permissions, startup type, and dependencies for reliable high availability behavior.

Regularly remove outdated clones to reduce storage cost and operational clutter. Old and unused clones cause high costs.

Maintain enough file system capacity above your total volume size. Too little free space causes performance issues; too much wastes storage and increases costs.

Configure thin provisioning for FSx for ONTAP volumes to make better use of storage and reduce costs. Thin provisioning lets you allocate storage as needed instead of reserving all capacity at once.

Turn on autosize and set mode to grow so volumes increase automatically before they become full and read-only.

Configure block device space settings for LUNs used by Oracle databases to prevent write failures and improve space efficiency on FSx for ONTAP.

Set snapshot copy reserve to 0% and enable snapshot autodelete to protect space for active data.

When space runs low, expand the volume before deleting snapshot copies. Expanding the volume prevents "volume full" errors, reduces write failures, and preserves snapshot recovery points.

Use the snapshot-only tiering policy, which moves only snapshot data to the capacity tier while keeping active data on the SSD tier. This approach preserves low-latency performance for SQL workloads while reducing storage costs. Tier snapshot data after a 7-day cooling period.

Enable deduplication, compression, and compaction to reduce storage use and cost while maintaining performance.

Set swap space based on the amount of RAM so the system can handle memory pressure and avoid slowdowns or crashes.

Place archive logs, data files, control files, redo logs, and temp files on separate drives to isolate I/O patterns and improve resiliency. Keep redundant copies on separate volumes where required.

Place Oracle binaries on a dedicated volume to reduce I/O contention. This separation simplifies software updates and minimizes the risk of accidental modifications or corruption.

Disable scheduled snapshots for FSx for ONTAP volumes used by Oracle to save space and lower costs. Instead, manage snapshots externally using tools such as NetApp SnapCenter, which create application-consistent backups and help prevent data corruption during restore operations.

Disable the ONTAP nfs-rootonly parameter for dNFS. nfs-rootonly restricts NFS connections to privileged ports (<1024). Because dNFS processes in NFSv4+ do not run as root and use higher ports, disabling this parameter allows necessary connections.

Make sure that if Oracle binaries are located on an NFS share, the export policy includes superuser and setuid permissions. Superuser (root) access allows NFS clients to map as root, needed for binary execution.

Use Oracle Automatic Storage Management (ASM) for iSCSI-based storage running on FSx for ONTAP to optimize performance, simplify storage management, and enhance scalability for Oracle Database deployments.

Configure the kernel parameters for the TCP slot table to 128 for Oracle workloads running on NFS. This configuration enables more concurrent I/O operations, reducing latency in high-throughput scenarios such as Oracle data file access. Kernel parameters govern the maximum number of outstanding NFS requests over TCP, acting like queue depth in Fibre Channel setups. The default Linux values, typically set at 16, do not support optimal database performance.

Use optimized NFS mount options for database files. Key parameters include rw for read/write access, bg for background mounting, hard for indefinite retries on failures, and proto=tcp for reliable delivery. The rsize and wsize values of 262144 improve large I/O operations, enhancing throughput, while nointr doesn't interrupt long-running tasks. This configuration improves database performance and resilience, particularly in high-throughput environments.

Use optimized NFS mount options for the Automatic Diagnostic Repository (ADR) home, or root directory. The ADR home stores logs and traces, which benefit from similar reliability options as core database files but don't require the nointr mount option because operations are less interrupt-sensitive.

Match NFSv4 domain names between the host (/etc/idmapd.conf or hostname -d) and NFS server (v4-id-domain in ONTAP).

Omit NFS mount options that disable host caching—such as, actimeo=0 and noac—in standalone Oracle deployments to improve efficiency and reliability. These options apply to clustered environments to ensure cache consistency, but in single-instance environments, they unnecessarily bypass caching and can significantly reduce performance.

Enable Direct NFS (dNFS) in your Oracle environment to improve NFS storage performance and reliability.

Avoid the use of Direct NFS (dNFS) with any type of round-robin name resolution, including DNS, DDNS, NIS, or any other method. This includes the DNS load balancing feature available in ONTAP. Consistent IP address resolution maintains database stability and prevents potential crashes or data corruption.

Check and update the oranfstab file content to ensure proper Direct NFS (dNFS) usage. The oranfstab file is essential for configuring advanced dNFS features such as multipathing and specific NFS options. Correct configuration improves the efficiency of data access and management.

Configure the nosharecache mount option for environments where Direct NFS (dNFS) is enabled, and a source volume is mounted more than once on a single server with nested NFS mounts. This configuration prevents cache sharing between mounts, ensuring data consistency and optimal performance.

Apply the latest operating system patches to improve security and reliability.

Disable Transparent Huge Pages (THP) on database hosts running Oracle databases to prevent potential performance degradation and improve memory management.

Enable TCP timestamps, SACK, and window scaling for best network performance and reliability.

Set filesystemio_options = setall for optimal I/O performance. Adjust the system global area (SGA) size if needed when moving away from buffered I/O.

Remove db_file_multiblock_read_count parameter from the text-based FILE init.ora or server-managed binary SPFILE (spfile<SID>.ora) for Oracle databases to prevent performance issues and allow Oracle to manage this setting automatically.

Apply the latest Oracle Critical Patch Updates to address security vulnerabilities and maintain the integrity of your database environment. Regularly review and apply patches to protect against potential threats and ensure compliance with security best practices.

Use application-consistent snapshots with NetApp SnapCenter to take accurate, reliable snapshots of your volume data at a specific moment in time. SnapCenter makes backups easier and helps you restore data quickly, reducing downtime and protecting critical workloads.

Enable cross-region replication (CRR) to improve data availability and disaster recovery for your Oracle databases. CRR helps protect against regional outages and ensures business continuity by replicating data across different locations.

Configure FSx for ONTAP backups or AWS Backup to meet retention and compliance requirements while avoiding duplicate backup workflows.

Regularly remove outdated clones to reduce storage cost and operational clutter. Old and unused clones cause high costs.