Standby databasesAn Oracle database can be configured with either a physical or a logical standby database for recovery after a production failure. At this time, Streamkap does not support them.
Oracle 12c or above, Standard or Enterprise Edition
Hostname and Port
(single-tenant architecture): Database name
(multi-tenant architecture): Container database name (CDB) and pluggable database name (PDB)
The Oracle database master user credentials or equivalent
For the Connector to ingest changes from your database it is dependent on Oracle’s redo logs and archive logs. It is important that these redo logs are large enough and, the archive logs are retained for long enough to ensure all changes are captured.
Reboot requiredWhen ARCHIVELOG mode is enabled, your Oracle database will be taken offline.
When redo logs fill up, Oracle archives groups of them into archive logs. Archive logs should be retained for at least 24 hours. However, we recommend retaining them for longer. Too short a retention period and changes may not be captured and processed.
Retention periods and database storageArchive logs are retained on your database instance using up its storage capacity. It is important to make sure it has enough space, otherwise, performance issues and outages can occur.In general, the more tables (and columns) there are, the more capacity is required. For the Connector, additional capacity is essential because it depends on Oracle’s supplemental logging.To estimate what storage capacity you might need and assuming your Oracle database has supplemental logging enabled, you can look at the last 1 hour of log storage usage and multiply that by the retention period (at least 24 hours) you choose. Here’s an example script for that:
SQL
Copy
Ask AI
SELECT SUM(BLOCKS * BLOCK_SIZE) bytes, -- usage SUM(BLOCKS * BLOCK_SIZE) * 72 estimated_bytes -- assuming 72 hours archivelog retention FROM V$ARCHIVED_LOG WHERE FIRST_TIME >= SYSDATE-(1/24) -- last hour AND DEST_ID=1;
You should set the DB_RECOVERY_FILE_DEST_SIZE parameter to a value that is appropriate for your available disk space.
Copy
Ask AI
-- Replace the {...} placeholders as requiredalter system set db_recovery_file_dest_size = {recovery_file_size};alter system set db_recovery_file_dest = '{recovery_file_location}' scope=spfile;
Then, configure Oracle RMAN to retain backups and archive logs for at least 24 hours. We recommend retaining data for longer.
Copy
Ask AI
RMAN> CONFIGURE RETENTION POLICY TO RECOVERY WINDOW OF 3 DAYS;
For the Connector to query the redo and archive logs, it is dependent on the Oracle LogMiner utility. To enable that, supplemental logging must be enabled.
After supplemental logging has been enabled at the database level, you then need to enable table level supplemental logging.
Copy
Ask AI
-- To enable all supplemental logging, run the following SQL statement for each table:ALTER TABLE {schema}.{table} ADD SUPPLEMENTAL LOG DATA (ALL) COLUMNS
To confirm if supplemental logging has been enabled, run this query:
Copy
Ask AI
SELECT NAME, SUPPLEMENTAL_LOG_DATA_MIN FROM V$DATABASE;
If the SUPPLEMENTAL_LOG_DATA_MIN is YES it is enabled.
An Oracle instance typically starts with three, online redo log files, 50MB each. That is too small and too few; more log files are necessary, and the logs should be resized to at least 1024MB or more, especially for production databases.Before making any changes, run this query to check the current log file sizes:
Copy
Ask AI
SELECT GROUP#, BYTES/1024/1024 SIZE_MB, STATUS FROM V$LOG ORDER BY 1;
GROUP#
SIZE_MB
STATUS
1
50
INACTIVE
2
50
INACTIVE
3
50
CURRENT
Also, we need to determine how many log files per group there are and their location.
Copy
Ask AI
SELECT GROUP#, LOCATION FROM V$LOGFILE ORDER BY 1, 2;
GROUP#
LOCATION
STATUS
1
/opt/oracle/oradata/ORCLCDB/redo01.log
ACTIVE
2
/opt/oracle/oradata/ORCLCDB/redo02.log
INACTIVE
3
/opt/oracle/oradata/ORCLCDB/redo03.log
UNUSED
Now replace the old log files with new, larger log files. Only INACTIVE and UNUSED groups can be dropped and recreated.
Log multiplexingOracle does support the notion of multiple log files per group known as ‘log multiplexing’. If your database uses this, use a comma-delimited list of filenames to register each log file.
Copy
Ask AI
-- Execute these statements as many times as required for existing INACTIVE and UNUSED log groups-- Replace {group_number} placeholder and {log_location}, using a comma separated list of filenames if database uses log multiplexingALTER DATABASE CLEAR LOGFILE GROUP {group_number};ALTER DATABASE DROP LOGFILE GROUP {group_number};ALTER DATABASE ADD LOGFILE GROUP {group_number} ('{log_location}') size 1024M REUSE;
Now switch the ACTIVE log so we can drop it.
Copy
Ask AI
ALTER SYSTEM SWITCH LOGFILE;
We need to wait for the database to eventually switch the status of the ACTIVE group to INACTIVE. The switch could take several minutes, so be patient and recheck the size periodically. Once the status reaches INACTIVE, replace the last, old log file and confirm all logs are resized:
Copy
Ask AI
-- Execute these statements as many times as required for existing INACTIVE and UNUSED log groups-- Replace {group_number} placeholder and {log_location}, using a comma separated list of filenames if database uses log multiplexingALTER DATABASE CLEAR LOGFILE GROUP {group_number};ALTER DATABASE DROP LOGFILE GROUP {group_number};ALTER DATABASE ADD LOGFILE GROUP {group_number} ('{log_location}') size 1024M REUSE;SELECT GROUP#, BYTES/1024/1024 SIZE_MB, STATUS FROM V$LOG ORDER BY 1;
GROUP#
SIZE_MB
STATUS
1
1024
CURRENT
2
1024
UNUSED
3
1024
ACTIVE
…
…
…
Now add additional, new log files as necessary. We suggest starting with 4 additional log files.
Group numbering and log file sizingThe {group_number} should follow sequentially. There is no benefit to organising log groups in sequences such as 10,20,30 or mixed sizing such as 128MB, 1024MB, 512MB.For example, if the highest group number is 3 and log file /opt/oracle/oradata/ORCLCDB/redo03.log of size 1024MB, the next log file added should be group 4 and log file /opt/oracle/oradata/ORCLCDB/redo04.log of size 1024MB.
Copy
Ask AI
-- Execute these statements as many times as required-- Replace {group_number} placeholder and {log_location}, using a comma separated list of filenames if database uses log multiplexingALTER DATABASE ADD LOGFILE GROUP {group_number} ('{log_location}') size 1024M;SELECT GROUP#, BYTES/1024/1024 SIZE_MB, STATUS FROM V$LOG ORDER BY 1;
Depending on your database architecture, choose the correct script below to create a database user with privileges for the Connector.
Copy
Ask AI
-- Replace {...} placeholders as neededALTER SESSION SET CONTAINER=CDB$ROOT;CREATE USER C##STREAMKAP_USER IDENTIFIED BY {password};ALTER SESSION SET CONTAINER={PDB};CREATE TABLESPACE STREAMKAP_LOGMINER_TBS DATAFILE {filename} SIZE 25M AUTOEXTEND ON MAXSIZE UNLIMITED;ALTER USER C##STREAMKAP_USER DEFAULT TABLESPACE STREAMKAP_LOGMINER_TBS;ALTER USER C##STREAMKAP_USER QUOTA UNLIMITED ON STREAMKAP_LOGMINER_TBS;-- Grant permissionsGRANT CREATE SESSION TO C##STREAMKAP_USER CONTAINER=ALL; GRANT SET CONTAINER TO C##STREAMKAP_USER CONTAINER=ALL; -- Allows the Connector to use LogMinerGRANT LOGMINING TO C##STREAMKAP_USER CONTAINER=ALL;-- Flashback queries used for performing initial snapshots of the dataGRANT FLASHBACK ANY TABLE TO C##STREAMKAP_USER CONTAINER=ALL; GRANT SELECT ANY TRANSACTION TO C##STREAMKAP_USER CONTAINER=ALL; -- Required for schema history when performing initial snapshotsGRANT SELECT_CATALOG_ROLE TO C##STREAMKAP_USER CONTAINER=ALL; GRANT EXECUTE_CATALOG_ROLE TO C##STREAMKAP_USER CONTAINER=ALL; -- Connector creates a table for explicitly managing the flushing of internal log buffers (LGWR)GRANT CREATE TABLE TO C##STREAMKAP_USER CONTAINER=ALL;GRANT CREATE SEQUENCE TO C##STREAMKAP_USER CONTAINER=ALL; -- Grant the Streamkap user permission to read each schema and table you wish to syncALTER SESSION SET CONTAINER={PDB};GRANT SELECT ON {schema}.{table} TO C##STREAMKAP_USER CONTAINER=ALL;-- Alternatively, you can grant access to all-- GRANT SELECT ANY TABLE TO C##STREAMKAP_USER CONTAINER=ALL;-- Grant the Streamkap user access to the DBA_EXTENTS, DBA_TABLESPACES, DBA_SEGMENTS, and TRANSACTION system views. GRANT SELECT ON DBA_EXTENTS TO C##STREAMKAP_USER;GRANT SELECT ON DBA_TABLESPACES TO C##STREAMKAP_USER;GRANT SELECT ON DBA_SEGMENTS TO C##STREAMKAP_USER;GRANT SELECT ANY TRANSACTION TO C##STREAMKAP_USER;-- Grant the Streamkap user permission to run LogMinerALTER SESSION SET CONTAINER=CDB$ROOT;GRANT SELECT ON SYS.V_$DATABASE TO C##STREAMKAP_USER;GRANT SELECT ON SYS.V_$PARAMETER TO C##STREAMKAP_USER;GRANT SELECT ON SYS.V_$ARCHIVED_LOG TO C##STREAMKAP_USER;GRANT SELECT ON SYS.V_$ARCHIVE_DEST TO C##STREAMKAP_USER;GRANT SELECT ON SYS.V_$LOGMNR_CONTENTS TO C##STREAMKAP_USER;GRANT EXECUTE ON DBMS_LOGMNR TO C##STREAMKAP_USER;GRANT EXECUTE ON DBMS_LOGMNR_D TO C##STREAMKAP_USER;GRANT SELECT ANY TRANSACTION TO C##STREAMKAP_USER;GRANT EXECUTE_CATALOG_ROLE TO C##STREAMKAP_USER;
To backfill your data, the Connector needs to be able to perform Snapshots (See Snapshots & Backfilling for more information). To enable this process, a table must be created for the Connector to use. Depending on your database architecture, choose the correct script below.
Please create the signal table with the name STREAMKAP_SIGNAL. It will not be recognised if given another name.
Copy
Ask AI
-- Replace {...} placeholders as neededALTER SESSION SET CONTAINER={PDB};-- Create a local (PDB) userCREATE USER STREAMKAP_USER IDENTIFIED BY {password};GRANT CREATE SESSION TO STREAMKAP_USER; GRANT SET CONTAINER TO STREAMKAP_USER;ALTER USER STREAMKAP_USER DEFAULT TABLESPACE STREAMKAP_LOGMINER_TBS;ALTER USER STREAMKAP_USER QUOTA UNLIMITED ON STREAMKAP_LOGMINER_TBS;-- Create the tableCREATE TABLE STREAMKAP_USER.STREAMKAP_SIGNAL ( id VARCHAR2(255) PRIMARY KEY, type VARCHAR2(32) NOT NULL, data VARCHAR2(2000) NULL);-- Grant necessary privileges on the table to the common userGRANT SELECT, UPDATE, INSERT, DELETE ON STREAMKAP_USER.STREAMKAP_SIGNAL TO C##STREAMKAP_USER;
Connectors use “offsets”—like bookmarks—to track their position in the database’s log or change stream. When no changes occur for long periods, these offsets may become outdated, and the Connector might lose its place or stop capturing changes.Heartbeats ensure the Connector stays active and continues capturing changes.There are two layers of heartbeat protection:
Layer 1: Connector heartbeats (enabled by default)
The Connector periodically emits heartbeat messages to an internal topic, even when no actual data changes are detected. This keeps offsets fresh and prevents staleness.No configuration is necessary for this layer; it is automatically enabled. We recommend keeping this layer enabled for all deployments.
Why we recommend configuring Layer 2While Layer 2 is crucial for low-traffic or intermittent databases, we recommend configuring it for all deployments. It provides additional resilience and helps prevent issues during periods of inactivity.
You can configure regular updates to a dedicated heartbeat table in the source database. This simulates activity, ensuring change events are generated consistently, maintaining log progress and providing additional resilience.How this layer is configured depends on the connection type (if supported by the Source):
Read-write connections (when Read only is No during Streamkap Setup): The Connector updates the heartbeat table directly.
Read-only connections (when Read only is Yes during Streamkap Setup): A scheduled job on the primary database updates the heartbeat table, and these changes replicate to the read replica for the Connector to consume.
This layer requires you to set up a heartbeat table—and for read-only connections, a scheduled job (e.g., pg_cron for PostgreSQL, event_scheduler for MySQL)—on your source database.
Read-write connections
For read-write connections (when Read only is No during Streamkap Setup), the Connector writes to the heartbeat table directly.
Copy
Ask AI
-- Replace {...} placeholders as neededALTER SESSION SET CONTAINER={PDB};-- Create the heartbeat table with id, text, and last_update fieldsCREATE TABLE STREAMKAP_USER.STREAMKAP_HEARTBEAT ( id NUMBER GENERATED BY DEFAULT ON NULL AS IDENTITY PRIMARY KEY, text VARCHAR2(4000), last_update TIMESTAMP DEFAULT CURRENT_TIMESTAMP);-- Grant permission to the Streamkap userGRANT SELECT, UPDATE, INSERT, DELETE ON STREAMKAP_USER.STREAMKAP_HEARTBEAT TO STREAMKAP_USER;-- Grant necessary privileges on the table to the common userGRANT SELECT, UPDATE, INSERT, DELETE ON STREAMKAP_USER.STREAMKAP_HEARTBEAT TO C##STREAMKAP_USER;-- Insert the first row into the heartbeat tableINSERT INTO STREAMKAP_USER.STREAMKAP_HEARTBEAT (text) VALUES ('test_heartbeat');
Connect via SSH Tunnel: The Connector will connect to an SSH server in your network which has access to your database. This is necessary if the Connector cannot connect directly to your database.
Username (case sensitive): Username to access the database. By default, Streamkap scripts use STREAMKAP_USER.
Password: Password to access the database.
Database: The database name (single-tenant architecture) or container database name (multi-tenant architecture).
Pluggable Database (optional, multi-tenant architecture only): The pluggable database name.
Heartbeats:
Heartbeat Table Schema: Streamkap will use a table in this schema to manage heartbeats. Usually this is the same as the Signal Table Schema. See Heartbeats for setup instructions.
Represent binary data as: Specifies how the data for binary columns should be interpreted. Your destination for this data can impact which option you choose. Default is bytes.
Capture Only Captured Databases DDL: Specifies whether the connector records schema structures from all logical databases in the database instance or only captured databases. Enabling this when you have many databases in your instance can improve performance and avoid timeouts. Default is false. See Schema History Optimization for details.
Capture Only Captured Tables DDL: Specifies whether the connector records schema structures from all logical tables in the captured schemas or databases, or only captured tables. Enabling this when you have many tables can improve performance and avoid timeouts. Default is false. See Schema History Optimization for details.
Add Schemas/Tables: Specify the schema(s) and table(s) for capture.
You can bulk upload here. The format is a simple list of schemas and tables, with each entry on a new row. Save as a .csv file without a header.
CDC only captures base tables, not ViewsChange Data Capture reads Oracle redo logs via LogMiner, which only record changes to physical tables. Database Views are query-time computations with no physical storage—they don’t generate redo log entries.What you cannot capture: Views, materialized views (capture base tables instead), global temporary tables, external tables, or system tables (DBA_*, ALL_*, USER_*).Solution: Specify only the underlying base tables that feed your views. You can recreate the view logic in your destination or transformation layer.
Click Save.
Have questions? See the Oracle Source FAQ for answers to common questions about Oracle sources, troubleshooting, and best practices.