Applying best practices to the server installation
Typically, hardware configuration and selection have the most significant effect on the performance of an IBM Storage Protect solution. Other factors that affect performance are the operating system selection and configuration, and the configuration of IBM Storage Protect.
Procedure
- The following best practices are the most important for optimal performance and problem prevention.
-
Review the table to determine the best practices that apply to your
environment.
Best practice More information Use fast disks for the server database. Enterprise-grade solid-state disks (SSD), with Fibre Channel or SAS interface, offer the best performance. Use fast, low-latency disks for the database. Using SSD is essential if you are using data deduplication and data replication. Avoid Serial Advanced Technology Attachment (SATA) and Parallel Advanced Technology Attachment (PATA) disks. For details and more tips, see the following topics: Planning for server database disks
Planning for the correct type of storage technology
Ensure that the server system has enough memory. Review operating system requirements in technote 84861. Heavier workloads require more than the minimum requirements. Advanced features such as data deduplication and data replication can require more than the minimum memory that is specified in the system requirements document. If you plan to run multiple instances, each instance requires the memory that is listed for one server. Multiply the memory for one server by the number of instances that are planned for the system.
Separate the server database, the active log, the archive log, and disk storage pools from each other. Keep all IBM Storage Protect storage resources on separate disks. Keep storage pool disks separate from the disks for the server database and logs. Storage pool operations can interfere with database operations when both are on the same disks. Ideally, the server database and logs are also separated from each other. For details and more tips, see the following topics: Planning for server database disks
Planning for server recovery log disks
Planning for storage pools in DISK or FILE device classes
Use at least four directories for the server database. For larger servers or servers that use advanced features, use eight directories. Place each directory on a LUN that is isolated from other LUNs and from other applications. A server is considered to be large if its database is larger than 2 TB or is expected to grow to that size. Use eight directories for such servers.
See
Planning for server database disks.
If you are using data deduplication, data replication, or both, follow the guidelines for database configuration and other items. Configure the server database according to the guidelines, because the database is extremely important to how well the server runs when these features are being used. For details and more tips, see the following topics: For storage pools that use FILE type device classes, follow the guidelines for the size of storage pool volumes. Typically, 50 GB volumes are best. Review the information in Optimal number and size of volumes for storage pools that use disk to help you to determine volume size.
Configure storage pool devices and file systems based on throughput requirements, not only on capacity requirements.
Isolate the storage devices that are used by IBM Storage Protect from other applications that have high I/O, and ensure that there is enough throughput to that storage.
For more details, see Checklist for storage pools on DISK or FILE.
Schedule IBM Storage Protect client operations and server maintenance activities to avoid or minimize overlap of operations. For more details, see the following topics: Monitor operations constantly. By monitoring, you can find problems early and more easily identify causes. Keep records of monitoring reports for up to a year to help you identify trends and plan for growth. See Monitoring and maintaining the environment for performance.