As of ReadyNAS OS 6.10, you can now add data tiers to Flex-RAID volumes. This adds onto existing functionality from OS 6.9 (metadata tiering). For more information about metadata tiering, see ReadyNAS OS 6.9: Metadata Tiering.
Data tiers are special RAID groups that primarily receive freshly written data onto a faster SSD RAID group, and eventually flush (or migrate) to spindle hard drive RAID groups. As such, data tiers provide faster writes overall compared to metadata tiering. Metadata tiering also takes place when using data tiering following the metadata rules notated at the article above. If you were previously using metadata tiering, your volume will continue to use metadata tiering until you manually switch the volume to data tiering.
Data tiers are recommended to have two or more SSDs. In RAID 0/JBOD volumes, the minimum number of SSDs is one. Two SSDs in a tier will act like a RAID 1. After expanding from a two-SSD tier to a three-SSD tier, your tier will act like a RAID 5. You can convert your data tier from RAID 5 to RAID 6 after three disks by adding parity.
Data tier storage is mostly temporary, depending on your configuration. Your data will migrate from SSD RAID groups to spindle RAID groups on a scheduled basis, when SSD-tier storage usage reaches a defined capacity, or when you choose to manually migrate the data.
Before adding a data tier to your volume, you need to consider your volume's rate of data change and your current metadata usage in order to plan accordingly. Minimally, your SSD data-tier capacity should exceed your daily, weekly, or monthly rate of newly written data to the volume, depending on how often you want data to flush to spindle RAID groups. Migrating data from the SSD-tier to the spindle-tier can result in heavy IO until the SSD-tier has fully migrated all stored data. As such, you may not want to exceed the defined capacity during peak business hours as this may affect volume performance. As metadata tiering also applies when data tiering, you need to ensure that your current and future metadata allocations, in addition to your expected rate of data change, do not exceed your SSD tier capacity.
On volumes with multiple data-tiered RAID groups, both metadata and data are written first to the tier with the most capacity available, if the tiers have differing available capacity levels. Once multiple tiers share the same available space, data and metadata will alternate between all available data tiers.
It is not recommended to mix different disk types in the same RAID group or tier.