The sizing requirements are the same the traditional star ETL process.
It is recommended to create all tablespaces as BIGFILE TABLESPACE
- STAR_DAT1
- STAR_HST1
- ODI_TGT_DAT1
- UNDOTBS1
Additionally, you will need to:
- UNDO tablespace size for staruser.
This may require unusual large size if spread data is huge. Usually it needs to hold UNDO for both activity and assignment spread whole tables inserts.
- Set ODI_TGT_USER tablespace size to hold all data for all data sources.
Tablespace | Files | Size |
---|---|---|
UNDOTBS1 | undotbs01.dbf | 32G |
ODI_TGT_DAT1 | ODI_TGT_DAT1.dbf | 32G |
| ODI_TGT_DAT2.dbf | 17G |
- Set STAR_DAT1 and STAR_HST1 tablespace.
- Ensure you plan ahead for adding extra data files, especially for STAR_HST1.
- Plan Processes in pfile or spfile
120 concurrent connections per P6 data source if all load plans run at the same time
- Concurrent connection needed by load plans:
- On Demand: 20
- Real time project: 20
- Real time global: 30
- Daily: 30
- Load single project: 20
- 30 concurrent connections per Unifier data source
For optimal values see Doc ID 240710.1 to decide after the ETL is stable for a period of time.
- Concurrent connection needed by load plans: