One FETCHER for Two MINE Processes

Problem Description

Can one FETCHER process handle multiple MINE processes?

Solution

Yes it is possible. Let us consider the below Setup:

• Source Database and Server: TESTDB(dbvlin801).

• Target Database and Server: DEVDB(dbvlin802).

• Target Database and Server: DEVTARG(dbvlin803).

• Replication Pairs:

  • TESTDB => DEVDB

  • TESTDB => DEVTARG

• DDC NAME: RTEST

We will configure FETCHER running on server dbvlin801 that will handle the MINE and MINE1 processes for the respective replication pairs. The MINE process will run on server dbr01, MINE1 on dbr02.

After completion of the Dbvisit Replicate setup wizard:

1. Enter the DDC directory. Example:

   cd /home/oracle/RTEST

2. Edit the following parameter in RTEST-onetime.ddc created in the config folder:

   SET FETCHER1.PROCESS_TYPE OFF SET FETCHER.MINE_PEER MINE:MINE1

   Note: We change FETCHER.MINE_PEER=MINE and FETCHER1.MINE_PEER=MINE1 after the setup wizard but before running Dbvisit Replicate.

3. Run the RTEST-all.sh script. It will show the message "WARN-9448: Cannot select process FETCHER1, it's type OFF." but this can be safely ignored.

4. Make sure data instantiation processes are complete before you proceed.

5. Edit:

   RTEST-run-dbvlin801.dbvisit.co.nz.sh

   and remove the FETCHER1 entry. Either delete or comment out the line:

6. We can now start the processes.

7. Start the console to follow progress.

Note: It will take some time for the APPLY process to catch up with MINE. Also you might need to adjust the memory parameters for both MINE and APPLY processes. For more information, please see:

https://dbvisit.atlassian.net/wiki/display/ugd7/Configuration+Variable+Reference

Ankit Arora April 20, 2015 21:38