Teradata Data Stream Architecture (DSA) supports backup and restore of Teradata Managed OTF tables using an in-place backup model. This model preserves table definitions while keeping data in your object storage.
In-Place Backup Overview
Unlike traditional backups that copy entire datasets, the in-place backup model:
- Archives only the dictionary metadata stored in Teradata’s Block File System (BFS).
- Sets protection markers in the Teradata data dictionary to prevent accidental table deletion.
- Does not archive any table data stored in your object storage.
Backup Strategy
- Backup Count Configuration: When backups are initiated on Teradata Managed OTF tables, DSA sets the backup count in the SQLE metadata to avoid inconsistencies between the backup set and the data stored in the object store.
- Default Backup Exclusions: In VantageCloud Lake, Backup-As-a-Service (BaaS) excludes Teradata Managed OTF tables from the default backup configuration. For these tables, use a partial backup strategy designed to meet their specific protection needs.
Disaster Recovery is not supported now.
The Apache Iceberg and Delta Lake read and write capabilities are enabled within Teradata Vantage. No special installation or setup is required to enable the feature. The user environment must comprise of an Iceberg or Delta Lake with the appropriate credentials and accessto the Iceberg/Delta Lake catalog and object storage.
Teradata OTF is a new feature that is automatically installed with Database Engine 20. Following is the installation and setup information for Teradata OTF:
- Customer OTF Iceberg or Delta Lake data must be stored on AWS S3-API compatible storage (for example, Dell ECS, AWS S3) or, for Azure, on Azure ADLS2 storage or on Google Cloud Storagefor GCP.
- Customers must provision a catalog service (such as Hive, Glue, or Unity) that can access the storage service.