Recovering table spaces or indexes to a point on the log

You can recover a single object, or a list of objects. You can recover an entire table space, index, a partition or data set. When you recover to a point on the log, RECOVER puts any associated index spaces in RECOVER pending status. Select to recover an index from the Control Center to remove the index space from RECOVER pending status.

If you recover a table space set, then you must ensure that you recover the entire set of table spaces, including all indexes on auxiliary tables to a common quiesce point. If you do not include every member of the set, or if you do not recover the entire set to the same point in time, then RECOVER puts all dependent table spaces of the inconsistent table spaces in the set in CHECK pending status.

If the table space or index space to be recovered is associated with a storage group, DB2 deletes and defines the necessary data sets. If you have used STOGROUP from the Control Center to remove the volume on which the table space or index space is located, RECOVER places the data set on another volume of the storage group.

Authorities and privileges

To recover a table space or index:

  1. Open the Recover To Logpoint window.

  2. In the Parallel processing options select a radio button to:

    1. Decide whether to process the table spaces or indexes in parallel.
    2. Let DB2 choose the maximum number of objects to process in parallel.
    3. Specify the number of table spaces or indexes to process in parallel. If you specify 0, or leave the field blank, DB2 chooses the maximum number of objects to process in parallel.

  3. Select the Specify action for data sets check box to:

    1. Select the LOGONLY check box if you have restored a data set to a point of consistency by using a data set copy that was not made by the DB2 utility COPY. This is because you do not want DB2 utility RECOVER to begin by restoring the data set from a DB2 image copy. Selecting this check box skips the RESTORE phase of RECOVER and applies the log records only, starting from the first log record written after the data set was backed up.
    2. Select the REUSE check box to place records in non-empty data sets.

  4. Select the Recover to a point on the log based on check box and specify a point on the log to recover to. The LOGPOINT is in hexadecimal format (X 'byte-string'). You can use a log record sequence number (LRSN) or a relative byte address (RBA). The LRSN is a number DB2 generates and associates with each log record. The LRSN is a string of 12 hexadecimal digits.

  5. Select the Use image copies from check box to tell the RECOVER to use image copies from the local or recovery site.

  6. Optional: Click the Add push button to add an object (table space or index).

  7. Optional: Click the Change push button to modify an object (table space or index).

  8. Optional: Click the Remove push button to delete a specified object from the page.

  9. Click OK to begin recovering.