Affinity sensitive storage of data corresponding to a doubly mapped redundant array of independent nodes

A method, comprising:

• determining, by a system comprising a processor and a memory, current affinity values for a current double affinity matrix corresponding to a first mapping of first mapped disks of mapped clusters to first real disks of a real cluster, wherein the current doubly affinity matrix reflects at least one current node affinity value and one current disk affinity value, wherein a node affinity value of the at least one current node affinity value corresponds to a count of mapped clusters that are supported by a first real node of the real cluster, and wherein a disk affinity value of the at least one current disk affinity value corresponds to a count of mapped clusters that are supported by a real disk of the first real disks;
• determining, by the system, first predicted affinity values for a first updated double affinity matrix corresponding to a second mapping of second mapped disks of the mapped clusters to second real disks of the real cluster, wherein the second mapping is a different mapping than the first mapping, wherein the updated doubly affinity matrix reflects at least one updated node affinity value and one updated disk affinity value;
• and
• in response to determining that a selection rule is satisfied based on the current affinity values and the first updated affinity values, allocating, by the system, the second real disks in accord with the second mapping.

A non-transitory machine-readable storage medium, comprising executable instructions that, when executed by a processor, facilitate performance of operations, comprising:

• receiving current node affinity values and current disk affinity values corresponding to a current mapping of first mapped disks of mapped clusters with first real disks of a real cluster, wherein a node affinity value of the current node affinity values corresponds to a count of mapped clusters that are supported by a first real node of the real cluster, and wherein a disk affinity value of the current disk affinity values corresponds to a count of mapped clusters that are supported by a real disk of the first real disks;
• receiving first predicted node affinity values and first predicted disk affinity values corresponding to a first alternate mapping of second mapped disks of the mapped clusters with second real disks of the real cluster;
• receiving second predicted node affinity values and second predicted disk affinity values corresponding to a second alternate mapping of third mapped disks of the mapped clusters with third real disks of the real cluster;
• in response to determining that a selection rule is satisfied based on the current node affinity values and current disk affinity values, the first predicted node affinity values and first predicted disk affinity values, and the second predicted node affinity values and second predicted disk affinity values, selecting a selected mapping, selected mapped disks corresponding to the selected mapping, and selected real disks corresponding to the selected mapping, wherein the selected mapping is selected from a group of mappings comprising the first alternate mapping and the second alternate mapping;
• and
• provisioning the selected mapped disks by allocating the selected real disks in accord with the selected mapping..