1.2.12 Datasets

An Exascale dataset is a logical grouping of files within a vault. The primary purpose of a dataset is to enable tracking and management of storage utilization for files associated with Oracle Database and Oracle Grid Infrastructure.

Exascale automatically creates and maintains system-defined datasets, which are conceptually organized in a hierarchical tree. Each dataset directly contains some number of files and indirectly contains child datasets that correspond with other entities in the hierarchy. The hierarchy of system-defined datasets contains the following levels:

1. Exascale vault

2. Oracle Grid Infrastructure (GI) cluster

3. Oracle multitenant container database (CDB)

4. Oracle pluggable database (PDB)

Starting at the bottom, the files belonging to each PDB are contained in a separate dataset. The parent CDB dataset contains the files belonging to the CDB and all of the associated PDB datasets. All of the CDBs in a GI cluster as grouped in a GI-level dataset, which also contains GI-specific files such as the Oracle Cluster Registry (OCR) and voting files. At the top of the hierarchy, the vault-level dataset contains all of the GI-level datasets that consume storage in the vault. The vault-level dataset is also the default container for any files that don't belong to Oracle Database and Oracle Grid Infrastructure.

Each system-defined dataset is identified by a composite ID, which contains unique identifiers for the associated entities in the hierarchy. For example, each vault-level dataset is simply identified by the vault name (for example, @MYVAULT). But, a PDB dataset ID has the following format:

@Vault-name:GI-cluster-ID:CDB-ID:PDB-ID

Each system-defined dataset also has a name. The dataset name is also a composite of the relevant entities, but it contains human-readable names for the GI cluster, CDB and PDB components (instead of the system-generated unique identifiers). For example, a PDB dataset name has the following format:

@Vault-name/GI-cluster-name:CDB-name.PDB-name