RESTORE keyword

The Recovery utility automatically performs a two-phase recovery for a specific DBD if it detects that the input image copy is an Instant Snapshot copy produced by the BMC Image Copy utility or a T0 copy produced by Data Base Image Copy 2 utility from IBM. For more information, see The-two-phase-recovery-process.

If the input data set is a T0 copy, the following considerations apply:

• The Recovery utility processes AREC statements synchronously (as if you had specified REC statements).
• For recovery of a Fast Path multiple area data set (MADS), the Recovery utility restores only the first area data set. The utility issues a warning message and ignores the others.

You can also use the RESTORE keyword to control task overlap when you are using extract data sets as follows:

• When you use the default value RESTORE(N) with the DIRECT(Y) keyword, the recovery tasks wait to start until the utility has written all extracted data to the extract data sets. In this approach, the utility writes output blocks only one time; it merges image copy, change accumulation, and log records in storage and then writes the merged records to the output database.
• If you specify RESTORE(Y) with the DIRECT(N) keyword, the utility begins downloading or restoring the image copies at the same time that it is creating the log extract data set and CA extract data set. When the restore is complete and the extract data sets are written, the utility rereads the extract data and applies them in a roll-forward mode (in which only the updated blocks are written). With this approach, more I/O is required to the output database; however, for large databases with a small percentage of updated blocks, this approach can provide a faster overall recovery.