Oracle Goldengate Introduction

Goldengate software is a log-based structured data replication software. GoldenGate can realize the real-time capture, transform and delivery of large amount of transaction data, synchronize the data between the source database and the target database, and maintain the sub-second data delay.

Goldengate can support multiple topologies, including one-to-many, multi-pair, cascade and bidirectional replication, etc.

There are several topological structures for data replication

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/73/26/wKioL1X2orPgktQFAAGjiRdjeoI136.jpg "title=" GG1. PNG "alt=" Wkiol1x2orpgktqfaagjirdjeoi136.jpg "/>

Goldegate's application Scenario

• A static extraction of data records from one database and the loading of those records

  to another database.

• Continuous extraction and replication of transactional DML operations and DDL

  Changes (for supported databases) to keep source and target data consistent.

• Extraction from a database and replication to a file outside the database.

Architecture Overview

Oracle Goldengate is made up of the following components

• Extract

• Data Pump

• Replicat

• Trails or Extract Files

• Checkpoints

• Manager

• Collector

650) this.width=650; "src=" Http://s3.51cto.com/wyfs02/M01/73/26/wKioL1X2pqOxeya7AAER5169rw0662.gif "title=" 20111116032004930.gif "alt=" Wkiol1x2pqoxeya7aaer5169rw0662.gif "/>

Extract process used to capture data sources, there are three types of

1. When configured as initial load mode, the data source is a table.

2. The recovery log for the database (some databases are also referred to as transaction logs).

3. A third-party capture model that sends data for database changes to the extract process by calling the Extact API.

Data pump is the auxiliary optional component of extract, and if data pump,extract is not configured, it is sent directly to the collector process on the target machine.

Using data pump can provide the following advantages

1. Prevent collector on the network or target, and protect data loss.

2. Data pump can be used to worry about, transform, or configure into pass-through mode, that is, to do nothing, just post data.

3. Launch multiple data pump to achieve a one-to-many distribution (or data from a single table to different destination systems depending on the data type)
   

Replicat runs on the destination server, reads the data from the trail file, reconstructs the DML, DDL statements, and applies to the destination database. You can configure multiple Replicat to work concurrently to increase the throughput of your system.

Trail in order to support the continuous capture and replication of database changes, Oracle captures the database changes temporarily to a series of disk files. These disk files are called Trail file.

A trail file can only be written by one extract. Each extract process must be linked to a trail file.

Checkpoints the Goldengate instance can be resumed when the goldengate process has been down.

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/73/27/wKioL1X2wp_itMO5AAL1uHQvHkI145.jpg "title=" gg2. PNG "alt=" Wkiol1x2wp_itmo5aal1uhqvhki145.jpg "/>

Manager is the Goldengate management process.

Before extract and Replicat start, the manager process must be started on each node.

The manager has the following functions

• Start Oracle GoldenGate Processes

• Start Dynamic Processes

• Maintain port numbers for processes

• Perform Trail Management

• Create event, error, and threshold reports

Collector, receives the data transmitted by the remote system and writes it to the trail file.

Group Overview

To differentiate multiple extract or replicat processes in a system, you need to define a processing group

A group includes the following:

• A process, Extract or Replicat

• It's parameter file

• It's checkpoint file

• Any other files associated with this process.
  

This article is from the "Ding Dong" blog, please be sure to keep this source http://lqding.blog.51cto.com/9123978/1694724

Oracle Goldengate Introduction