Abstract: An index recovery mechanism recovers all indexes that were journaled and may be fully recovered during IPL. In addition, the index recovery mechanism performs a first stage of processing for any index that was journaled but cannot be fully recovered during IPL. The first stage processing includes storing index recovery data for an index during IPL. Once IPL is complete, the index recovery mechanism performs a second stage of processing by reading the index recovery data that was stored during IPL, and completing recovery of the index. By performing this two-stage index recovery, the first stage during IPL and the second stage post-IPL, indexes that cannot be fully recovered during IPL can still be journaled and recovered.

Abstract: According to a preferred embodiment, a method of sorting a list of elements with duplicate entries using multiple processors is disclosed. Using “P” processors, a list of elements is split into P lists and each processor pre-sorts a list. All pre-sorted lists are lined up to form a partitioning table, with each pre-sorted list making up a column in the table, and the first element from each pre-sorted list making up the first row in the table, and the second element from each pre-sorted list making up the second row, etc. P−1 partition boundary lines are drawn through the partition table to create P equally sized partitions. Each partition boundary line is drawn such that every element below the line has a value larger than any element above the line, and every element above the line has a value smaller than any element below the line. Duplicate elements are uniquely “weighted” during the partitioning process.