SQL/MX 3.2.1 Management Manual (H06.26+, J06.15+)

NOTE: The format of the TMF audit records for SQL/MX objects changed in SQL/MX Release
3.0. As a result, the R2.3.x systems cannot apply the new format (R3.x) audit information whereas,
R3.x systems can apply R2.3.x audit information. The R2.3.x primary systems can replicate audit
trail content to R3.x disaster recovery systems. This allows you to upgrade the backup system to
R3.x and test the upgrade prior to production. When the R3.x system becomes the active node,
the constraints are:
• R2.3.x systems cannot handle the R3.x audit trail content and therefore cannot act as a disaster
recovery site for the R3.x system.
• R2.3.x systems must be upgraded to R3.x before starting the RDF replication from R3.x systems.
Using Guardian Physical Names for SQL/MX Objects
For more information about using the MXGNAMES utility with RDF to facilitate the use of Guardian
physical names for SQL/MX objects, see “Using Guardian Names with TMF, RDF, and Measure”
(page 309).
Configuring Disk Volumes for RDF
You perform disk volume configuration for RDF as you do for SQL/MX databases. HP recommends
that volume names match exactly on the primary and backup nodes. If they do not, they can be
remapped for RDF with the RDF SET VOLUME configuration command.
Creating RDF Backup Catalogs
Create RDF backup catalogs with different names from their corresponding primary node catalogs
to enable:
• Certain database creation and population features to work with the catalog distribution feature.
• Migration to future versions of NonStop SQL/MX that might involve optional catalog-to-catalog
replication.
You cannot define the primary or backup catalogs as remote catalogs on their respective nodes.
That is, you should create primary catalogs on the primary node and register them on the backup
node. Conversely, create backup catalogs on the backup node and register them on the primary
node to ensure that each automatic catalog reference exists on the same nodes as its corresponding
database.
For example, if the primary catalog is named CAT, you can create the backup catalog as CATB:
CREATE CATALOG CATB [LOCATION $DATA03];
Execute this statement on the RDF backup node. The optional location clause specifies that you
place the metadata files on volume $DATA03.
Creating RDF Backup Schemas
SQL/MX schemas are created with a designated subvolume name where all Guardian files
associated with objects in that schema are placed. This subvolume name is stored in the
SCHEMATA.SCHEMA_SUBVOLUME column in the system schema entry for each user schema.
For example, if the node name is NODE1, this query returns the schema subvolume name for a
schema called CAT.SCH1:
SELECT S.schema_subvolume
FROM nonstop_sqlmx_node1.system_schema.schemata S,
nonstop_sqlmx_node1.system_schema.catsys C
WHERE S.schema_name = 'SCH1' AND C.cat_name = 'CAT' AND
S.cat_uid = C.cat_uid;
Using RDF to Set Up a Backup System 57