Is anybody using SORTNPSI AUTO in their Reorgs and if so have you noticed any benefit / difference?
I tried using it on one of our biggest tables. This is in a PBR, with 1.4 billion rows, 120 partitions and 3 indexes.
The Reorg uses PARTLEVEL and LISTPARTS 20, so effectively it ran with 6 individual Reorgs. However, each Reorg issued message DSNU1242I, indicating that they all sorted all keys and each of the 6 Reorgs took about 30 minutes longer than when using SORTNPSI NO and used more CPU.
I thought the idea behind SORTNPSI AUTO was that it could / should improve things. What am I missing?!