Timestamp recovery
Timestamp recovery is fully supported in the Recovery utility.
You can perform a normal timestamp recovery, a point-in-time recovery, or a special timestamp recovery.
When you perform a timestamp or point-in-time recovery, you should also recover any logically related or application related databases that were updated since the timestamp.
Always create an image copy after executing a timestamp or point-in-time recovery. Use the IC or STKIC keyword to generate the required image copy. BMC recommends that you take an image copy because the change accumulation process does not recognize that a database has been recovered. If you do not take an image copy after you recover a database and you execute a change accumulation task, the change accumulation task uses logs that were created before and after the database was recovered. However, if you need to perform a subsequent recovery and the database is registered to DBRC, DBRC ignores the second change accumulation data set and uses logs as input to the recovery.
To rebuild the primary index of a PHIDAM HALDB after a timestamp recovery, you must specify the BLDINDEX(Y) keyword, or you must run the DFSPREC0 utility. This action turns off the RECOVERY NEEDED flag for the ILDS component.
To rebuild the indirect list data set (ILDS) component of a HALDB after a timestamp recovery, you must specify the BLDILDS(Y) or BLDILDS(R) keyword, or you must run the DFSPREC0 utility, even if no indirect list entries (ILEs) are present. This action turns off the RECOVERY NEEDED flag for the ILDS component.
This section contains the following topics:
Related topic