1. Deploy a Directory-Based Distributed Database

A directory-based configuration follows the same steps as you would for a user-defined distributed database, with a few differences.

Most of the information you need is found in Oracle Globally Distributed Database Deployment for planning, installing and creating the databases for the distributed database topology.

To configure the topology for Directory-based sharding, do the following tasks:

1. Create a shard catalog for user-defined sharding. See Creating a Shard Catalog for Directory-Based Sharding.

2. Add and start shard directors. See Add and Start Shard Directors.

3. Create shardspaces, and shards in those shardspaces. See Add Shardspaces If Needed, Add the Shard CDBs, and Add the Shard PDBs.

4. Create tablespaces in the shardspaces. See User-Defined Data Distribution for examples.

   Note that each tablespace has to be created individually, and explicitly associated with a shardspace.

5. Verify the topology, add shards and host metadata, deploy the configuration, and start global database services. See Oracle Globally Distributed Database Deployment.

2. Create Schema Objects

• To create tables sharded by directory, see Creating Tables Sharded by Directory.

• To add (and remove) keys to the directory, see Managing Keys in Directory-Based Sharding.

3. Run DML and Queries

See DML Support on Tables Sharded by Directory.

4. Perform Lifecycle Operations

Over the lifetime of your directory-based distributed database, you'll need to do tasks such as:

• Add and remove keys. See Managing Keys in Directory-Based Sharding

• Add partitions. See Adding a New Tablespace and Chunks (Partition) in a Shardspace

• Move chunks. See Chunk Management in Directory Based Sharding.

• Split chunks. See Splitting Partitions (Chunks).

• Query the directory view for metadata. See Sharding Key Directory Public View.