From Goldengate Capture |verified| | Ogg Capture Client Successfully Detached

When an Extract process runs in mode on an Oracle database (versions 11.2.0.4 and later), it registers with the database as a "capture client." The database spawns a background process (typically a child of the LogMiner process or a dedicated V$GG_* process) that feeds transaction data to the Extract.

BEGIN DBMS_CAPTURE_ADM.STOP_CAPTURE('GG_CAPTURE_CLIENT_NAME'); DBMS_CAPTURE_ADM.DROP_CAPTURE('GG_CAPTURE_CLIENT_NAME'); END; /

The Extract process died abnormally (e.g., kill -9 , network split, storage loss) before sending the "detach" command. The database kernel still believes the client is alive. When an Extract process runs in mode on

This message typically appears in logs when a Extract process (capture client) is gracefully detached from its capture source (e.g., redo logs, trail files, or a coordinated capture). It signals a clean, expected shutdown or detachment of the capture component from the main capture mechanism.

This message typically occurs during a standard shutdown of an Extract process. In Integrated Capture This message typically appears in logs when a

If the Extract refuses to start even after database cleanup, the local purgestamp file may be corrupt. Use:

), it must "detach" to release resources, such as memory and hung sessions in the LogMiner engine. Key Observations Normal Behavior In Integrated Capture If the Extract refuses to

cd $GG_HOME/dirchk rm E<extract_name>.pce

During diagnostic operations like FETCHLOG or specific ALTER EXTRACT, SCN commands, the Extract may temporarily attach to a specific SCN range, read it, and then detach upon completion.