SAP Replication Server Notes for Oracle Sources – Architecture, Use Cases, Versions, Known Restrictions and Limitations, and Performance Tuning Settings

Today’s enterprise supports a complex, ever growing data environment that often includes incompatible systems and data schemas, spread across the internet – LAN, WAN, and the cloud. With SAP® Replication Server®, your data is always where you need it to be. This post is intended to consolidate and present some useful Oracle source related information.

System Architectures

From Oracle to HANA: Multiple components are required to implement an Oracle-to-SAP HANA database replication system: a primary Oracle data server, a replicate SAP HANA database, a Replication Server with ExpressConnect for HANA DB, and Replication Agents for Oracle.

Typical Oracle to HANA Replication Use Cases

1. Data off-loading for Reporting and Analytics
2. Data Migration

From Oracle to Oracle: Multiple components are required to implement an Oracle-to-Oracle database replication system: a primary Oracle data server, a replicate Oracle database, a Replication Server with ExpressConnect for Oracle DB, and Replication Agents for Oracle.

Typical Oracle to Oracle Replication Use Cases

1. Data off-loading for Reporting
2. Data Migration
3. Disaster Recovery

Examples of Supported Versions

Note: We support Exadata sources (http://help.sap.com/Download/Multimedia/zip-rso1571204/Replication_Server_Options_New_Features_Guide_en.pdf).

Known Restrictions and Limitations

Replication Agent for Oracle: http://infocenter.sybase.com/help/index.jsp?topic=/com.sybase.infocenter.dc73463.1571200/doc/html/aga1269937573549.html

ExpressConnect for Oracle:

http://infocenter.sybase.com/help/index.jsp?topic=/com.sybase.infocenter.dc73463.1571200/doc/html/bde1276905201353.html

Key Performance Tuning Settings

# This Replication Agent setting turns on parallel scanning.

# Set this to true if you need the performance.

# Default value is false.

ra_config lr_parallel_scan, true

go

# This Replication Agent setting turns on filtering.

# This setting will allow row and column level filtering via where clauses in replication definitions

# Default value is true.

ra_config lr_parallel_scan_with_filter, true

go

# This Replication Agent setting sets the number of parallel scanners for parallelism (range 1-64).

# 5 is a reasonable suggested starting number for parallelism.

# This setting can be tuned based on need and Replication Agent node resource capacity

# Default is 4

ra_config lr_parallel_scanner_count, 5

go

# This Replication Agent setting sets the number of SCNs (Sequence Numbers) per scanner (range 128 to 2147483647)

# 1000 is a reasonable suggested starting range if you need parallelism.

# This setting can be tuned based on need and Replication Agent node resource capacity

# Default is 1024

ra_config lr_parallel_scan_range, 1000

go

# This Replication Agent setting sets the of records queued in the scanner queues

# 10000 is a reasonable suggested starting number.

# This setting can be tuned based on need and Replication Agent node resource capacity

ra_config lr_parallel_scan_range, 100000

go

# This Replication Agent setting sets the extra SCN (Sequence Number) range to check the end of each scanner.

# This setting is useful for when you have LOBs and wide records.

# This will allow user to go as wide as they can until performance degrades.

# This setting can be tuned based on need and Replication Agent node resource capacity

# Default value is 1.

ra_config lr_parallel_scan_range_margin, 20

go

# This Replication Agent setting sets the queue size of each scanner extra SCN (size ranges from 0-2147483647)

# This setting is valid when lr_parallel_scan is set to true.

# The default setting is 0, that is unlimited size.

ra_config lr_parallel_scan_queue_size, 0

go

Related Resources

Data Replication Architecture for Integration Heterogeneous Systems with Oracle Databases

Heterogeneous Replication Guide

Replication Agent for Oracle