Maintaining recovery-related records affected by online reorganization

For recovery purposes, the A-through-J and M-through-V data sets are treated in the RECON data set as one logical set of DBDSs.

The recovery information in the two DBDS headers is maintained as if they are one logical DBDS. For example, the values for GENMAX, RECOVPD, and CAGRP are the same. Any changes that are made to the active DBDS are also made to the inactive DBDS.

There is a one-to-one relationship between each partition data set; for example, A correlates to M, B correlates to N, C correlates to O, and so on. Also, the M-through-V data sets are defined in the RECON data set with the same attributes as their corresponding A-through-J data sets. You can perform online reorganization on any of these data sets (A-through-J, L and X or M-through-V and Y) at any time, depending on which set is active before the online reorganization starts. Therefore, data sets A and M are one logical data set during an online reorganization.

After an image copy is taken, DBRC removes any extraneous recovery-related records (IMAGE, ALLOC, RECOV, and REORG) from the RECON data set. The extraneous records are determined by the maximum number of recovery generations (GENMAX) settings and the optional recovery period (RECOVPD) settings for the DBDS.

For example, assume that the GENMAX value for data set A (and M) is 2, and that data set A has two image copies. When data set M becomes the active data set and an image copy is taken, the image copy cleanup process deletes the oldest image copy from data set A, including any of its extraneous recovery records. Another image copy of data set M causes the rest of data set A's image copy and recovery records to be deleted. Note that data set A becomes unrecoverable if its image copy and recovery records are deleted.