NonStop SQL DDL Replicator User's Guide (Update 7)

SDR Operations
HP NonStop SQL DDL Replicator User’s Guide—545799-007
4-10
Replication Operations
replicated. It simply monitors RDF and performs the associated DDL operation when
RDF enters the “Update NSA Stopped” state.
When RDF enters the Update NSA Stopped state, the SDR updater accesses the RDF
backup copy of the depot file where the associated original DDL statement was stored.
Note that the SRU file name will have a subvolume beginning with ZASDR.
If the SRU file subvolume does not begin with ZASDR, then the audit was not
produced by SDR and SDR will ignore it. It may have been caused by a WITH
SHARED ACCESS operation occurring when SDR DDL capture is disabled. You must
deal with this situation using standard RDF (i.e., manual) methods.
Translating DDL
When the DDL has been fetched from the depot file, it is translated for execution on
the backup system. SDR uses the RDF configuration to replace file names and catalog
names with their RDF backup equivalents. SDR uses the RDF volume table, any
includes and excludes, and the mapfile if configured.
During the translation process, SDR distinguishes between the target SQL object and
any secondary file reference. The target object is the first file name in the DDL
statement, specifying the table or index or view being created or updated. All other
names, such as catalogs, partitions, physical volumes, a CREATE LIKE table, or a
table being indexed are secondary references.
If the DDL target volume is not replicated, or the target file name or subvolume is
excluded, SDR skips the operation and restarts the RDF updaters. If RDF replicates
the target name, but does not replicate some secondary reference, SDR signals an
error and waits for operator intervention.
Appendix A, SQL DDL Statements
provides information on how each DDL statement is
translated.
Executing DDL
When the DDL has been successfully translated, SDR executes it on the backup
database.
If the command drops a SQL object, then SDR refers to the REPLICATEPURGE
setting of RDF. If the setting is ON, the DROP is executed; otherwise SDR issues an
EMS message and awaits operator intervention.
Note. Many RDF configurations have excluded SQL catalog subvolumes. If your DDL
references a catalog on an excluded subvolume of a replicated volume, SDR will still attempt
to replicate the operation. It will use the same subvolume on the RDF backup volume.
But, if the backup catalog does not exist, SDR (and RDF) will wait for operator intervention.
You will have to either create the catalog manually or CANCEL the DDL replication.
This rule does not apply to CREATE CATALOG or DROP CATALOG on a subvolume that is
excluded in RDF. Also, in this situation, AUTOCREATECATALOG will not create a catalog.