Preparing for a failover of the CDC Replication Engine for Oracle databases
To prepare for a failover such as a hardware or software failure, your clustering environment will require a script that performs CDC Replication tasks on the new active node.
About this task
The tasks in the following steps must be included in your failover script.
Procedure
- Clean the transaction queues for each instance by removing all files that begin with txqueue* in the <CDC replication installation directory>/instance/<instance name>/txnstore directory.
- Back up your metadata for each instance by archiving all files that begin with md* in the <CDC replication installation directory>/instance/<instance name>/conf directory. If the metadata database does not recover after a failover, restore these files to the same directory on the new active node. This task is optional.
- Back up your Event Log messages for each instance by archiving all files in the <CDC replication installation directory>/instance/<instance name>/events directory. If the events database does not recover after a failover, restore these files to the same directory on the new active node. This task is optional.
- Start each instance on the new active
node with the following command:
dmts64 - I <instance_name> - Run the dmclearstagingstore command on the new active node for all instances.
- Start all subscriptions on the new active node with the dmstartmirror command.