StarQuest Technical Documents

SQDR Plus for iSeries Failover Procedures

Last Update: 10 March 2010
Product: SQDR Plus for iSeries
Version: 3.62 or later
Article ID: SQV00PL006

Abstract

Under certain circumstances, it may be necessary to move the SQDR Plus source database from a production system to a backup system (i.e., "fail-over" scenario) and subsequently to restore SQDR operations on a primary server (i.e., "fall-back" scenario). In both cases, SQDR Plus provides functionality to ease this transition.

This document describes the proper procedures to migrate from one source database to another, without requiring that SQDR run baseline replications for the incremental subscriptions. It is necessary that the version of SQDR Plus on the backup system be exactly the same as that running on the primary server. Contact StarQuest Customer Support if you no longer have the installation source media for your version.

These instructions assume that the SQDR client will remain on the same system and that no changes will be made to its configuration. If you need to migrate the SQDR client in tandem, follow the instructions in the StarQuest Technical Document titled Migrating SQDR Plus Subscriptions from a Test Environment to a Production Environment.

Before You Begin

On the Production iSeries

1. Ensure that the source tables will not receive any further updates. If necessary, end all non-SQDR related remote communication with the database.
2. Verify that the SQDR Plus Capture Agent has processed all of the transactions in the logs and staged all changed records, and confirm that the SQDR client has retrieved all changed data. On the DB2 source database, use CAMAINT, option 1 Display configured SQDR clients, to confirm that there are zero transactions (i.e., NumTx field is blank) pending for the groups/subscriptions.

On the SQDR client platform

1. Launch the Data Replicator Manager.
2. Monitor the incremental replication groups, the row count as displayed in the group statistics should not be incrementing at this point because the SQDR client should have processed all data. If the row count is still increasing, wait until all of the groups have finished replicating their change data.
3. Right-click each incremental replication group and select Pause Updates.
4. Stop the SQDR service.

Procedures

On the Production iSeries

1. Start the Capture Agent Maintenance Utility (CAMAINT).
2. Stop the Capture Agent using option 6.
3. Select the Recovery option to display the Agent Recovery Menu. Select option 1 Set Startup Mode and set Capture Agent start up mode to “Warm” startup with fully recovered control and staging tables. Exit the menu.
4. Save the SQDR library including all objects.

On the Backup iSeries

1. Install SQDR Plus.
2. End the SQDR subsystem in a controlled fashion.
3. Restore the SQDR library excluding any files of type *PKG.
4. In the SQDR library, delete the SQ_JRNMAP file and all files starting with SQ_JO0000<N> (where N is a sequential number, such as SQ_JO00001, SQ_JO00002, etc.)
5. Using CAMAINT, enter the Recovery menu and verify that the Capture Agent startup mode is "Startup with fully recovered control and staging tables". Exit the menu and exit CAMAINT.
6. Start SQDR Plus subsystem.
7. Verify operations have been restored by reviewing the Capture Agent and Log Reader status using CAMAINT.

On the SQDR Client

1. Launch the Data Replicator Manager, which will automatically start the SQDR service.
2. Resume Updates for the IR group(s).

At a later point in time, revert the Capture Agent startup mode back to Normal.

Refer to the SQDR Plus Documentation Addendum for more information about the Capture Agent Startup modes.