How does the Fetcher work with redo logs?

Problem Description

This article explains how fetcher process works.

Solution

The Fetcher process allows the Mine process to be run on a different server to the source database server. The Fetcher process runs on the source database server and sends the online redo logs from the database server to the Mine server. 

The MINE_STAGING_DIR directory contains copies of online/archive logs shipped from the source database server to the Mine server. This is done by fetcher sending the data to mine. Fetcher sends the data as soon as it is available, i.e. it ships online redo logs as they are written. Mine evaluates which redologs are no longer needed and deletes them, so the mine_staging_dir directory should not grow indefinitely.

The Mine process created plog files from the redo or archive redo logs files received from the Fetcher process.

The plogs are the files containing the information mined from the redologs. They are shipped (again, as soon as there is anything to ship, i.e. they are shipped even before they are finished) from mine to apply. Thus apply has a copy of the same files mine does have. Both mine and apply evaluate which plogs are no longer needed and deletes them.

Arjen Visser July 14, 2013 22:28