Declustered array size

IBM Storage Scale RAID distinguishes between large and small declustered arrays. A declustered array is considered large if, at the time of its creation, the number of its pdisks is at least the setting of the vcdSpares parameter + 9. When using the default values for the vcdSpares and dataSpares parameters, this means that a declustered array is considered large if it contains at least 11 pdisks. All other declustered arrays are considered small. At least one declustered array in each recovery group must be large, because only large declustered arrays have enough pdisks to safely store an adequate number of replicas of the IBM Storage Scale RAID configuration data for the recovery group.

Because the narrowest RAID code that IBM Storage Scale RAID supports for user data is 3-way replication, the smallest possible declustered array contains four pdisks, including the minimum required equivalent spare space of one disk. The RAID code width of the intended vdisk NSDs and the amount of equivalent spare space also affect declustered array size; if Reed-Solomon 8 + 3p vdisks, which have a code width of 11, are required, and two disks of equivalent spare space is also required, the declustered array must have at least 13 member pdisks. Declustered arrays that contain only log vdisks can be smaller than these limits.