Depending on which cluster manager software that is being used, you need to configure it to ensure the following:

• There is exactly one active node that is running Oracle GoldenGate for Distributed Applications and Analytics (GG for DAA). It is assumed that the cluster manager can detect that a compute node is down and subsequently spawn another node to be the active.
• Install GG for DAA in shared file system and have that shared file system mounted in the same location for all the nodes participating in the High Availability (HA) configuration. For more information about installing Oracle GoldenGate for Distributed Applications and Analytics (GG for DAA), see Installing Oracle GoldenGate for Distributed Applications and Analytics . Most of the state files, including the Input and Output Trail files, Configuration files, and Checkpoint files described in the next point are stored in sub-directories of the Oracle GoldenGate for Distributed Applications and Analytics (GG for DAA) install. The GG for DAA installation directory is the same across all managed nodes. This helps the administrator to leverage the exact content of entry point script to bring up GG for DAA as part of its workflow to spawn a new active node. An example of the content of the entry point script is a command to start the Oracle GoldenGate manager process.
• Oracle GoldenGate artifacts are stored in one or more shared file systems or volumes accessible from all nodes. For more information about these files, see Directories and Variables in Microservices Architecture:
  • Input and Output Trail files: Typically these files are located in the gg_install_dir/dirdat directory, where gg_install_dir is the Oracle GoldenGate installation directory, such as C:/ggs on Windows or /home/user/ggs on UNIX. These files are configurable.
  • Configuration files: The configuration files are located in the gg_install_dir/dirprm directory.
  • Checkpoint files: These files are stored in an internal subdirectory, such as the gg_install_dir/dirchk directory.
  • When using File Writer features, for example, File Writer handler, ADW, or Redshift integration, the file writer output files and the state files must be on shared volumes.

For more information about configuring cluster high availability for handlers, see Configuring Cluster High Availability.