Configure Advanced Settings

Explore different advanced settings for GoldenGate Studio.

You can use Advanced Options to fine tune your replications. These settings allow you to optimize performance, customize deployment behavior, and manage environment-specific requirements.

Initial Load Advanced Options

When setting up a replication, you can configure advanced options to optimize the Initial Load process.

The Initial Load advanced options determine how the data is transferred, handled, and synchronized between the source and target systems. When configuring Initial Load options in Oracle GoldenGate Studio, available settings depend on the selected database type. This section outlines the available options for Oracle and MySQL databases.

Oracle AI Database: Supported Initial Load Advanced Options

Initial Load options you can configure for Oracle AI Database are as follows:

Action Upon Existing Tables: Determines how to handle existing target tables during initial load. Expected values: REPLACE, TRUNCATE, APPEND, SKIP.
Degree of Parallelism: Number of parallel threads for the Data Pump load. Higher values can speed up loads but increase resource usage. If the number of CPUs is ‘n’ and degree of parallelism value is less than or equal to ‘n’ then the degree value is going to be ‘n’. However, if the number of CPUs are ‘n’ but the degree of parallelism value is ‘x’, which is greater than ‘n’, then the degree value is going to be ‘x’. Expected value: Any positive integer value greater than or equal to 1.
Additional Initial Load (Data Pump) Job Duration: Specifies for how much time jobs will run after the expected completion time for Initial Load. After the assigned time it will automatically time out. Expected value: 1h.
Transfer Medium: Specifies the transfer method for Initial Load. Expected values: Database Link, Object Storage, or File.
Object Storage Bucket URI: Location of the staging bucket for Initial Load files. Expected value: https://objectstorage.<region>.oraclecloud.com/.
Source Wallet URI: URI or path to the SSL Wallet for source database (required). Expected value: /u02/app/oracle/admin/sourcedb/ssl_wallet.
Target Database SSL Wallet Path: This is Wallet directory location of SSL Wallet for target database. For more information, see Configure and Download SSL Wallet for Non-Autonomous AI Database. Expected value: /u02/app/oracle/admin/targetdb/wallet.
Source Database SSL Wallet Path: This is Wallet directory location of SSL Wallet for source database. For more information, see Configure and Download SSL Wallet for Non-Autonomous AI Database. Expected value: /u02/app/oracle/admin/sourcedb/ssl_wal.
Wait Time for Open Transactions: Duration to wait for open transactions to finish before replication starts. Expected value: 1h.
Export Directory (Local Shared Storage): This is Source shared directory location for writing dump files during data pump export. Expected value: /mnt/source_exports.
Import Directory (Local Shared Storage): This is target shared directory location for writing dump files during data pump export. Expected value: /mnt/target_imports.
Actions Upon Expiry: Specifies if there is need to continue or stop if the transactions are open. Expected value: CONTINUE, STOP.

MySQL Database: Supported Initial Load Advanced Options

Initial Load options you can configure for MySQL are as follows:

Action Upon Existing Tables: Determines how to handle existing target tables during initial load. Expected values: SKIP, ERROR.
Degree of Parallelism: Number of parallel threads for the Data Pump load. Higher values can speed up loads but increase resource usage. If the number of CPUs is ‘n’ and degree of parallelism value is less than or equal to ‘n’ then the degree value is going to be ‘n’. However, if the number of CPUs are ‘n’ but the degree of parallelism value is ‘x’, which is greater than ‘n’, then the degree value is going to be ‘x’. Expected values: Any positive integer value greater than or equal to 1.
Transfer Medium: Specifies the transfer method for Initial Load. Expected value: Object Storage, or File Storage.

Note: GoldenGate Studio doesn’t support Database link for MySQL Database.
Object Storage Bucket URI: Location of the staging bucket for Initial Load files. Expected value: https://objectstorage.<region>.oraclecloud.com/.
Initial Load Dump Directory: Store dump files generated during Initial Load process. Dump directories are created under the ORACLE_BASE location for Oracle, and under the Studio installation directory for MySQL. Expected values: db_dumps (by default).

Note: This option applies when File Storage is selected as Transfer Medium.
Compression: Reduces the size of dump files. Expected values: NONE, GZIP, ZSTD.

Note: To get the Wallet Bucket URL from Oracle Cloud Infrastructure (OCI), where Oracle Autonomous AI Database is running, create a bucket in OCI and generate a pre-authenticated read/write URL for that bucket.

Data Compression Options

Learn different data compression options GoldenGate Studio offers.

The following data compression options are available in Configuration Advanced Settings for MySQL:

NONE

Applies no compression. Selecting this option keeps files at their original size, enabling the fastest processing speeds but resulting in higher storage and bandwidth usage.
GNU Zip (GZIP)

GZIP is one of the most established and widely used compression algorithms. Provides moderate compression rates and is widely supported across platforms. Selecting this option results in reduced file sizes and ensures compatibility with most tools and systems but may offer slower processing speeds compared to ZSTD. Use GZIP when you need maximum compatibility across tools and platforms, minimal CPU usage, or are working with smaller datasets where compression speed is not the main concern.
Zstandard (ZSTD)

Zstandard (ZSTD) is a modern compression algorithm designed for both high compression ratios and rapid processing speeds. Offers higher compression efficiency and faster compression/decompression speeds than GZIP. Prefer ZSTD for large datasets, modern data pipelines, or performance-sensitive applications. It may not be supported by all legacy or third-party systems.

Type of Initial Load Transfer Mediums

In GoldenGate Studio, the Initial Load process is used to transfer a complete dataset from the source to the target before starting ongoing replication. This ensures both databases are synchronized before Change Data Capture (CDC) begins. GoldenGate Studio offers different transfer mediums for performing the initial load. These transfer mediums are:

Database Link

The Database Link method transfers data directly from the source database to the target database using an established database link, without any intermediate storage. This method works best when both databases are in the same network or connected through a stable, high speed link. Since it uses SQL INSERT statements to move data, the process is straightforward but can be slower for large datasets due to SQL execution overhead. Database Link is less suited for large datasets or cross-region migrations, where latency could significantly slow down the process. It is best used for small to medium sized transfers where simplicity outweighs performance concerns.
Object Storage

The Object Storage method stages initial load data in an Oracle Object Storage before importing it into the target database. GoldenGate Studio first exports data from the source into Data Pump files, stores them in the specified bucket, and then imports them into the target. This approach eliminates the need for a direct network connection between source and target, which is suitable for cross region, cloud to cloud, or hybrid cloud migrations.

A direct database connection may be slow or restricted, so the data is staged in Object Storage and securely fetched by the target system. This method is highly reliable for large datasets and supports resumable transfers in case of interruptions. However, it requires additional configuration for bucket creation and wallet authentication, as well as extra storage space for staging files. It is best used when the migration involves large volumes of data across different regions or environments.
File Storage

The File Storage method facilitates initial load by exporting source data into disk files on the GoldenGate Studio server. These files are then ingested into the target database. This method is particularly effective when both source and target databases are non Autonomous AI Database and can access a shared storage layer, such as:
- Network File System (NFS) mounts
- Oracle File Storage Service (FSS)
- Docker volume mapping (for containerized database installations)
By leveraging shared storage, whether on premises or cloud based organizations can achieve high throughput data transfer and maintain control over the staging environment.

This method mirrors the Object Storage approach but is tailored for environments without access to native cloud object storage. By using shared storage, organizations can replicate the reliability and efficiency of object storage in on premises or hybrid setups.

GoldenGate Studio writes extracted data files to the shared location, and the target database reads these files for ingestion. This intermediate staging allows for data validation, auditing, and performance tuning before final load. It is particularly useful in secure environments with strict compliance requirements and high speed LAN connectivity.

This setup ensures fast, secure data transfer, audit-friendly staging, and full control over sensitive financial data.

To learn more about initial load processing, refer to the Precise Instantiation for Oracle Using Initial Load Extract and About Data Replication Components in GoldenGate Studio topics in GoldenGate Studio Microservices Documentation.

Initial Load Support per Database

Different databases support different options for performing the initial load.

The following table outlines the supported initial load methods for Oracle and MySQL source databases.

Initial Load Option	Oracle	MySQL
DBLink	Yes	No
File Storage	Yes	Yes
Object Storage	Yes	Yes

Support Matrix - Oracle AI Database Variations and Initial Load Methods

Learn about the supported Initial Load methods for different Oracle AI Database with recommendations.

The following matrix outlines the supported Initial Load methods in GoldenGate Studio for various Oracle AI Database environments, including Oracle Autonomous AI Database and non Autonomous AI Database. These methods applies to all Recipes, including One-way, Active-active, and ZeroETL Recipes. Non Autonomous AI Database will contain both On-premise and DBaaS (Database as a service which is in Oracle Cloud Infrastructure).

Oracle Databases	Data Pump via Database Link	Data Pump using File Storage	Data Pump using Object Storage	Recommendation
Non-Autonomous AI Database to non-Autonomous AI Database	Yes	Yes	Yes	Choose Database Link for small datasets; Object Storage for large migrations.
Autonomous AI Database to non-Autonomous AI Database	Yes	No	Yes	Use Object Storage for secure and reliable transfer.
Non-Autonomous AI Database to Autonomous AI Database	Yes (Private Network)	No	Yes	Prefer Object Storage; use Database Link only if private network connectivity is available.
Autonomous AI Database to Autonomous AI Database	Yes	No	Yes	Object Storage is recommended for cross-region or large dataset transfers.

Extract Advanced Options

Explore the Extract Advanced Options for Oracle and MySQL database.

Source Database timezone: Specifies the time zone of the source database. This is a critical setting for ensuring accurate timestamp replication. Expected values: EST,PST,UTC,IST.
Additional Extract parameters: This parameter instructs the Replicat process to ignore the records that encounter error and continue processing. These are custom parameters that can be used to handle specific error codes and conditions. Expected values: REPERROR (PROCEDURE, DISCARD).

If you select Extract auto restart the following options appear:

Max Retries: The maximum number of times the Extract process will attempt to restart after a failure, before it gives up. Expected values: 0,1,2,3,4,5.
Retry Delay: The time interval, in seconds, between each restart attempt. Expected values: 1s,2s,3s,4s.
Restart Window: The time frame within which the Max Retries count is applied. Expected values: 1m,2m,3m.

Note: The Restart on failure only option is available in this release. When enabled, the process restarts only after a failure, not for normal stops.
Failures: The total number of times the Extract task has failed within the specified Restart Window. Expected values: 0,1,2,3,4,5.

Note: The Disable task after retries exhausted option is available in this release. When enabled, the task is automatically disabled if the maximum retry attempts are reached without success.

To enable Autostart for Extract, use the following parameters:

extractAutostartEnable: Enables auto start for extract process. Default value: false.
extractAutostartDelay: Delay before starting extract process. Default value: 5s.

Replicat Advanced Options

Explore the Replicat Advanced Options.

Action upon DML Error: Specifies the action to take when a Data Manipulation Language (DML) operation fails. Expected values: DISCARD, IGNORE, KILL, DEFAULT(RETRY_OPERATION), ABORT_TRANSACTION.
Maximum Retry Count: The maximum number of tries can be specified by the users. Expected values: 0, 1, 2, 3, 4, 5.
Action upon DDL Error: Specifies the action to take when a Data Definition Language (DDL) operation fails. Expected values: DISCARD, IGNORE, KILL.
Delay for aborting the transaction: The amount of time the system will wait before forcefully aborting open transactions that have not completed. Expected values: 1h.
Additional Replicat Parameters: This parameter instructs the Replicat process to ignore records that encounter Oracle error and continue processing. These are custom parameters that can be used to handle specific error codes and conditions. Expected values: REPERROR (26961, DISCARD).

If Replicat Auto Restart option is enabled, you will see the following options:

Retry Delay: The amount of time (in sec) to pause between discovering that the process has terminated abnormally and restarting the process. Expected values: 0, 1, 2, 3, 4, 5.
Failure: The number of times a task or process has failed in the current monitoring window. Expected values: 0, 1, 2, 3, 4, 5.
Max Retries: The maximum number of times the Replicat process will attempt to restart after a failure. This setting works in conjunction with RETRYDELAY. Expected values: 0, 1, 2, 3, 4, 5.
Restart Windows: Defines the time frame within which the Max Retries count is applied. Expected values: 1m, 2m, 3m.

To configure auto start for Replicat, you can use the following parameters:

replicatAutostartEnable: Enables auto start for the Replicat process. Default value: false.
replicatAutostartDelay: Delay before starting the Replicat process. Default value: 5s.