-
When sizing an array:
- Assume that all I/O will be random. Assuming random I/O yields the best results.
- Before releasing any storage system to production, use Dell Live Optics on a simulated production system during at least a 24-hour period that includes the peak workload. The simulation helps define the I/O requirements. It might also be possible to use an I/O testing tool to simulate the production system. (See Table 3 for a sample of available I/O testing tools.) After production begins, repeat the analysis on the production system.
- Understand what level of ASM disk group redundancy (external, normal, high) is being considered. Using ASM external redundancy with PowerStore is recommended unless Extended Distance Oracle RAC Clusters are used. For Oracle RAC Extended Distance Clusters, use either normal or high ASM redundancy. To avoid the complexity of configuring Oracle RAC Extended distance clusters and licensing costs for Oracle RAC, consider using of PowerStore metro node Metro instead. For more information, see References.
- Have a good understanding of the application workloads (OLTP, OLAP, or hybrid).
- NVMe NVRAM drives are reserved for system write cache in PowerStore 1200T through PowerStore 9200T models. User and system metadata is written to the other drive types (NVMe SCM, NVMe SSD, SAS SSD). NVMe drives provide lower latency and higher bandwidth than SAS SSD, so are a great choice for an Oracle solution.
- Understand the required performance metrics of the servers connected to the PowerStore array. The IOPS and throughput help determine the number of disks required in the array, and throughput helps define the paths between the PowerStore array and the server.