This topic describes how to plan recovery resources.
Planning recovery resources
Build a plan for grouping copies on tape, grouping objects on change accumulation files, and using copies for all data sets of an object that allows NGT Recover to use all resources effectively.
NGT Recover will attempt to:
Use shared resources in order (copies including all data sets in data set number order; input change accumulation files in order by database name, table space name, and data set number; input stacked copies by file sequence number)
Group use of shared resource without having intervening steps that use other resources
NGT Recover must process all data sets of a partitioned table space together when a nonpartitioned index is being rebuilt or when multiple partitions are unloaded.
When change accumulation files are being created, the batch process must process the objects in the grouping order: database name, table space name, and data set number. Similarly, NGT Recover cannot alter a stacking order for output copies that is given by your JCL.
Combining copies for multiple objects with change accumulation for groups
If you are making a copy of all data sets of a space and also making change accumulation files for the same space, make the change accumulation file for all data sets in the space.
If you are stacking copies and building change accumulation files for groups of objects, you should coordinate the objects in the change accumulation group with the stacked copies. Include all of the stacked objects or a subset in the change accumulation group. Stack the copies in alphabetical order by database name and table space name with indexes included as indicated in Using change accumulation files for multiple objects.
Handling large partitioned objects with separate change accumulation files and copies
If you have very large table spaces that are partitioned, make specific data set (partition) copies and do not stack them with other partitions or other table spaces.
Create a change accumulation group for each partition if the change accumulation file will be tape or if a large numbers of updates will occur between image copies.
If the change accumulation file will be disk, and there are not a large number of updates between copies, it may be appropriate to have all of the partitions grouped into the same change accumulation group. This plan allows a straightforward recovery of any partition or concurrent jobs on the partitions.