Replication using Streams at RLS o Oracle Streams

Replication using Streams at RLS o Oracle Streams n new way to share information between databases n Data replication: n o captures DML and DDL changes (events - LRC) o propagates events o applies changes at destination databases requirements: DB version (>=9. 2. 0. 4) o DB must run in archive log mode o o RLS database n OS: Linux n initial DB version: 9. 2. 0. 5 n capture staging consumption upgrade to 10 g (10. 1. 0. 2. 0)

Replication using Streams at RLS o Streams configuration n two machines located at CERN n streams setup in both directions n schema level replication: two schemas data copy RLS 1 streams data copy APPLICATION RLS 1 R 1

Replication using Streams at RLS o RLS stress test n written in Python; multi-thread application; producer-consumer pattern n uses full RLS application stack of POOL File. Catalog maintains consistent lists of accessible files (physical and logical names) together with their unique identifier (File. ID) resolves a logical file reference (File. ID) to a physical file n allows to setup multiple writers and readers using one or multiple RLS end-points which write and/or read RLS data via RLS end-points n log file: can be analyzed to produce some statistics n designed to avoid conflicts – not allow conflict resolution studies

Replication using Streams at RLS o Stress test run configurations n one end-point r RLS 1 rite n two end-points w n about 50. 000 entries n during 3 weeks RLS 1 r ir te w r ir te der w rea streams RLS 1 R 1 r rite w r e ad re r ir teader w re o success results o conflicts can occur => necessary look at handling conflicts

Replication using Streams at RLS o Next tasks: n n n conflict handling monitoring add external replication sites RLS 1 streams CNAF streams ……. RLS 1 R 1