Aggregates and RAID groups
Modern RAID technologies protect against disk failure by rebuilding a failed disk’s data on a spare disk. The system compares index information on a “parity disk” with data on the remaining healthy disks to reconstruct the missing data, all without downtime or a significant performance cost.
An aggregate consists of one or more RAID groups. The RAID type of the aggregate determines the number of parity disks in the RAID group and the number of simultaneous disk failures the RAID configuration protects against.
The default RAID type, RAID-DP (RAID-double parity), requires two parity disks per RAID group and protects against data loss in the event of two disks failing at the same time. For RAID-DP, the recommended RAID group size is between 12 and 20 HDDs and between 20 and 28 SSDs.
You can spread out the overhead cost of parity disks by creating RAID groups at the higher end of the sizing recommendation. This is especially the case for SSDs, which are much more reliable than capacity drives. For HDD aggregates, you should balance the need to maximize disk storage against countervailing factors like the longer rebuild time required for larger RAID groups.