How to increase WAFL cache memory by using FlexScale

You can increase Write Anywhere File Layout (WAFL) cache memory in storage systems running Data ONTAP by using FlexScale, which includes the Performance Accelerator Module as the hardware component and the WAFL extended cache as the software component. FlexScale needs to be licensed.

If you use FlexScale on active/active storage systems, you must ensure that FlexScale options are the same on both nodes. Otherwise, a takeover can result in lower performance due to the lack of FlexScale on the remaining cluster node.

• FlexScale limits
  FlexScale increases the performance of the storage system by reducing disk reads. With FlexScale's Performance Accelerator Module and its WAFL extended cache, your storage system's cache memory can hold up to 127 aggregates (which would mean 127 traditional volumes, if your system uses only traditional volumes). As long as you have fewer than 127 aggregates total, there is no limit to the number of flexible volumes that can be cached by the WAFL extended cache.
• Enabling and disabling FlexScale
  You can enable the WAFL extended cache functionality for a storage system that has the Performance Accelerator Module installed.
• Caching normal user data blocks
  If you cache normal user data blocks, the WAFL extended cache interprets this setting as the buffer cache policy of keep and saves normal user data blocks in the extended cache.
• Caching low-priority user data blocks
  You can cache low-priority user data blocks that are not normally stored by FlexScale. Low priority blocks include blocks read in large sequential scans that are not normally reused, and blocks that have been written to the storage system through a network-attached storage (NAS) protocol such as Network File System (NFS). Caching low-priority user data blocks is useful if you have workloads that fit within FlexScale memory and the workloads consist of either write followed by read or large sequential reads.
• Caching only system metadata
  If the working set of the storage system is very large, such as a large E-mail server, you can cache only system metadata in FlexScale memory by turning off both normal user data block caching and low-priority user data block caching.
• Integrating FlexShare buffer cache policies with FlexScale cache options
  For additional cache control, you can integrate FlexShare buffer cache policies with the FlexScale cache options (flexscale.normal_data_blocks and flexscale.lopri_data_blocks). You can choose to enable only the FlexShare buffer cache policies without enabling all other FlexShare options.
• Displaying FlexScale configuration
  Data ONTAP enables you to display information about FlexScale configuration.
• Displaying FlexScale usage and access information
  You can display usage and access information for FlexScale, have output produced periodically, and terminate the output after a specified number of iterations.