Oracle Golden Gate software is a log-based structured data replication backup software. It obtains incremental changes of data by parsing the online log or archived log of the source database, and then applies these changes to the target database, thereby realizing the source database and the target database. Synchronize. Oracle Golden Gate can achieve sub-second real-time replication of large amounts of data between heterogeneous IT infrastructures (including almost all common operating system platforms and database platforms), so that it can be used in emergency systems, online reports, and real-time data warehouses. , transaction tracking, data synchronization, centralization/distribution, disaster recovery, database upgrade and migration, dual business centers and other scenarios. At the same time, Oracle Golden Gate can implement a variety of flexible topologies such as one-to-one, broadcast (one-to-many), aggregation (many-to-one), bidirectional, point-to-point, and cascade.
1. OGG is data synchronization software, or logical replication software, its competitors are IBM CDC, DELL shareplex, DSG, DDS and so on.
2. The OGG software needs to be installed on the operating system, and the Oracle DB software does not have its own OGG software (at least for now) 3. The
OGG software has three major processes (or four major processes):
Three major processes: source extraction Process, source transmission process, destination application process (ie: copy process)
four processes: source extraction process, source transmission process, destination server process (responsible for writing trail files at the destination), destination application process (ie: Copy process)
4. The OGG software has two trail files
from the trail file source: written by the source extraction process.
Trail file of the destination end: The source transmission process transmits to the destination end, and the server process of the destination end writes it to the specified path of the destination end operating system.
5. The three major processes in the OGG software have their own checkpoints (ie: checkpoints, which are used to record where each process has read and where it has been written)
Source extraction process: There are read checkpoints and write checkpoints
Source transmission process: There are read checkpoints and write checkpoints
Destination application process: There are read checkpoints
6. The three major processes of OGG software process the trail file process :
Source extraction process: read the oracle redo log (assuming the source database is Oracle), combine the parameter file of the extraction process, change the content of the redo log into a file in OGG's own format (ie: trail file), and write it to The source transmission process under the specified path of the source operating system
: read the "trail file written by the source extraction process", combine the parameter file of the transmission process, and send the data that needs to be synchronized to the destination, where the server process is responsible for writing .
Destination application process: Read the trail file in the specified path of the destination operating system, combine with the parameter file of the application process, form SQL statements, and execute these SQL statements on the destination db.
GoldenGate mainly includes Manager process, Extract process, Pump process, and Replicat process. The following describes them one by one:
- The Manager process is the control process of GoldenGate, running on the source and target ends. Its main functions are as follows: start, monitor, restart other processes of Goldengate, report errors and events, allocate data storage space, publish threshold reports, etc. There is only one manager process on the target side and the source side
- Extract runs on the database source and is responsible for capturing data from source data tables or logs. The role of Extract can be divided into stages according to:
- Initial time loading stage: In the initial data loading stage, the Extract process directly extracts data from the data table on the source side
- Synchronization change capture phase: After the initial data synchronization is completed, the Extract process is responsible for capturing the changes in the source data (DML and DDL)
- The Data Pump process runs on the database source end, and its function is to send the local trail file generated by the source end to the target end in the form of data blocks through TCP/IP protocol, which is usually the recommended way. The pump process is essentially a special form of the extract process. If the trail file is not used, the extract process will directly deliver the data to the target end after extracting the data to generate a remote trail file.
- The Collector process and the Data Pump process correspond to the Server Collector process. This process does not require my attention, because in the actual operation process, we do not need to configure it, so it is transparent to us. It runs on the target side, and its task is to reassemble the data delivered by Extract/Pump into a remote ttrail file.
- Replicat process, usually we also call it the application process. Running on the target side, it is the last stop of data transfer, and is responsible for reading the content in the target side trail file, parsing it into DML or DDL statements, and then applying it to the target database.