Failback refers to the process of restoring normal user and client access to a primary database, after a failover switched access from the primary database to a standby database.
This section describes only the failback tasks that are specific to the Mirror Activator system. General system failback procedures are not covered in this document.
Table B-2 provides a checklist of the tasks that you must include in failback procedures for the Mirror Activator system.
The checklist in Table B-2 assumes that:
You have purged the Replication Server stable queue to remove any open transactions remaining after failover.
The primary data server is up and functioning correctly, and you are ready to return it to normal operation.
If you do not purge the Replication Server stable queue
to remove any open transaction that remained in the queue after
failover:
Replication will not start, because Replication Server will not send any new transactions to the standby database until it receives operations to complete the open transactions.
Eventually, the stable queue will run out of space, because Replication Server will not send any transactions after you resume replication upon failback.
Sybase recommends that you perform these tasks in the order shown.
Task |
Description |
|---|---|
1 |
Quiesce the standby database to suspend update activity. |
2 |
Fail back the disk replication system to:
|
3 |
Bring the primary database online. |
4 |
Initialize the primary database using the pdb_init move_truncpt command to set the truncation point to the end of the log. |
5 |
Quiesce the primary database. |
6 |
Initialize the Mirror Replication Agent using the ra_init force command, and set the paths to the mirror log devices, if necessary. |
7 |
Release the quiesce on the primary database, after the Mirror Replication Agent initialization is complete. |
8 |
Switch access for users and client applications from the standby database to the primary database. |
9 |
Release the quiesce on the standby database, after client access is switched to the primary database. |
10 |
Start replication in the Mirror Replication Agent with the resume command. |
Use the following procedure for Mirror Activator system failback tasks.
Failing back the Mirror Activator system
Quiesce the standby database to suspend update activity.
Fail back the disk replication system to:
Materialize the primary database data and log devices from the standby database devices.
Materialize the mirror log devices at the standby site from the materialized primary log devices.
Re-establish synchronous replication from the primary log devices to the mirror log devices at the standby site.
Refer to the documentation provided by your disk replication system vendor for more information about the procedures in this step.
For more information about re-materializing a primary database, see Appendix A, “Materializing Databases in a Mirror Activator system.”
Bring the primary database online.
Initialize the primary database using the Mirror Replication Agent pdb_init move_truncpt command. This command sets the truncation point to the end of the log.
For more information about initializing the primary database, see “Setting up the Mirror Activator system”.
Quiesce the primary database.
Initialize the Mirror Replication Agent using the ra_init force command to update the system data repository in the RASD, and then use ra_devicepath to set the paths to the mirror log devices (if necessary).
See “Updating the RASD”.
Release quiesce on the primary database, after the Mirror Replication Agent initialization is complete.
Switch access for users and client applications from the standby database to the primary database.
Release quiesce on the standby database, after client access is switched to the primary database.
Start replication in the Mirror Replication Agent.
Log in to the Mirror Replication Agent administration port and execute the following command:
resume
After you invoke resume, use the ra_status command to verify that the Mirror Replication Agent is in Replicating state.
The following output indicates that Mirror Replication Agent has reached the Replicating state:
State Action ----------- ------------------------------ REPLICATING Waiting for operator command.
