SQL/MX 3.2 Management Manual (H06.25+, J06.14+)

If CAT_1 and CAT_2 are tightly interrelated and if most application programs and queries access
objects in both catalogs, the configuration described in Table 23 (page 274) would instead look
like the configuration shown in Table 24 (page 275).
Table 24 Consequences of Network Node Loss, Reconfiguration 3
Consequences of Losing This Node on the Other Nodes
CAT_1 and
CAT_2
CAT_1 and
CAT_2
CAT_1 and
CAT_2 User
Node
Applications
Present?
Metadata
Present?
Data
Present?
From nodes \B and \C, you cannot perform DDL operations
on objects partitioned across node \A. You cannot perform
Yes, accesses
CAT_1 and
CAT_2
Yes for
CAT_1
Yes\A
DDL, DML, or utility operations on objects whose metadata
resides on node \A, including CAT_1 and CAT_2 objects
and programs. CAT_2 objects not partitioned across node
\A can be accessed.
From nodes \A and \C, you cannot perform DDL operations
on objects partitioned across node \B. You cannot perform
Yes, accesses
CAT_1 and
CAT_2
Yes for
CAT_2
Yes\B
DDL, DML, or utility operations on objects whose metadata
resides on node \B, including CAT_2 and CAT_1 objects
and programs. CAT_1 objects not partitioned across node
\B can be accessed.
From nodes \A and \B, you cannot perform DDL operations
on objects partitioned across node \C. You cannot perform
NoNoYes\C
DDL, DML, or utility operations on objects whose metadata
resides on node \C.
In this example, confining CAT_1 and CAT_2 metadata and the applications that query it to two
nodes (\A and \B) creates a dependency on the availability of both of those nodes for all operations
except the execution of statically compiled queries. This configuration is not recommended.
Creating a Distributed SQL/MX Database
You create a distributed SQL/MX database by:
• “Creating, Registering, and Unregistering Catalog References” (page 275)
• “Distributing SQL/MX Database Objects” (page 276)
For instructions on creating a distributed SQL/MP database, see the SQL/MP Installation and
Management Guide.
Creating, Registering, and Unregistering Catalog References
After a catalog has been created on the local node, use the REGISTER CATALOG command to
register the empty catalog reference on remote nodes and the UNREGISTER CATALOG to remove
an empty catalog reference from a node.
Example—Creating a Catalog
The CREATE CATALOG statement creates a new catalog on the designated node:
CREATE CATALOG samdbcat LOCATION \nodex.$data47;
Example—Registering a Catalog on a Remote Node
The REGISTER CATALOG command causes a catalog, which is visible on the local node, to be
registered on a remote node. It creates an empty catalog reference on the target node, and the
automatic catalog reference is updated to reflect that.
REGISTER CATALOG samdbcat ON \nodey.$data47;
Creating a Distributed SQL/MX Database 275