Capture from Oracle RAC

Fivetran HVR supports capturing changes from Oracle Real Application Clusters (RAC). Capture from Oracle RAC is supported using the Direct Redo Access, BFILE, or LogMiner capture method.

When capturing from Oracle RAC, HVR will typically connect with its own protocol to an HVR Agent Listener installed in the RAC nodes. The HVR Agent Listener should be configured to run inside all cluster nodes using the same Port, User/Login, and Password for all the Nodes. The hub then connects to one of the remote nodes by first interacting with the Oracle RAC 'SCAN' address.

HVR requires that the agent property Agent_Server_Public_Certificate across all RAC nodes be the same. To achieve this, before creating an Oracle RAC location, you must copy the HVR Agent configuration files (hvragent.user and hvragent.conf available in HVR_CONFIG/etc directory) from the primary RAC node to the rest of the nodes.

• The hvragent.user file is available only if you have created agent users.
• If the Agent_Server_Public_Certificate across all RAC nodes are not same, you may encounter the F_JG2194: Remote certificate does not match required certificate error.

The HVR channel only needs one location for the RAC and there is only one capture job at runtime. This capture job connects to just one node and keeps reading changes from the shared redo archives for all nodes.

Directory HVR_HOME and HVR_CONFIG should exist on both machines but does not normally need to be shared. If HVR_TMP is defined, then it should not be shared.

In HVR UI, the connection parameters for the Oracle RAC must be supplied while creating an Oracle location with Oracle RAC in Step 2. Select Agent Connection.

When capturing changes from an Oracle RAC, if you define the Capture Checkpoint Tuning option, the checkpoint files should be stored on the hub server. This is because the directory on the remote location where the capture job would otherwise write checkpoints may not be shared inside the RAC cluster. Consequently, it may not be available when the capture job restarts. For more information, see Tuning Capture Checkpoints section.