How the Snapshot Copy function works

If a supported intelligent storage device is not available, the Snapshot Copy function uses the software cache. In this case, the XBM subsystem intercepts any updates made to a database that is registered for SUF usage and preserves the preimages in cache. (Preimages are copies of database blocks before they are updated; cache is virtual storage obtained by the XBM subsystem.) The Snapshot Copy function includes the preimages in the Snapshot Copy, not the corresponding updated blocks that are read from the database. Including the preimages in the Snapshot Copy ensures that the Snapshot Copy is a true batch image copy, reflecting an image of the database at the start of the image copy, even though the database is updated by application programs during the copy process.

Software-assisted Snapshot Copy processing flow

To deallocate databases that are allocated when Snapshot Copy processing begins, use the AOI. For more information, see Automated-operator-interface.

To start Snapshot Copy processing

1. Register databases with XBM. (Any XBM registration failure equates to a card-scan or control statement failure, and no further Snapshot Copy processing takes place.)
2. If CONSIST(GLOBAL) is specified, get DBRC authorization for the entire consistency group.

3. If necessary, issue a /DBR or /DBD command for full-function databases or a /STO command for Fast Path databases.

   Important

   When DBQUISCE(Y) is selected, both the Instant Snapshot and the Snapshot Copy functions are not supported.

4. If CONSIST(GLOBAL) is specified, get the timestamp for the point of consistency.
5. Start the Snapshot Copy process.
6. Release DBRC authorization.
7. Issue a /STA command to start the databases where applicable.
8. Copy the databases.
9. Stop the Snapshot Copy process.