Configuring Async Node Restarts after Database Failures

In the event of a database failure, by default the async grid node will attempt to restart indefinitely. That is, assuming the database is restarted, the async node should eventually also restart without manual intervention. With this procedure, you can change the restart behavior. You have three choices:

  • Keep the restart retry period indefinite.

  • Set the restart retry period to a specific number of hours.

  • Remove the automatic restart. In this case, the async node restart would need to be done manually.

If the restart retry period is set to indefinite or to a specific number of hours, the async node will sleep 10 seconds for the first 10 tries, 1 minute for the next 10, 10 minutes for the next 10, then an hour at a time after that. Before sleeping, it will check to see if it has exceeded the maximum time and will stop attempting to restart if it has.

This feature is meant to deal with unintended database shutdowns. If you are shutting down the database on purpose, we recommend that you first run stoplaw. After you have restarted the database, you then would run startlaw in order to start all of the grid nodes, including the async node.

To configure async node restarts

  1. In the Grid Management pages, access the page for your Landmark application.
  2. Select Configuration > Edit Properties > Service Execution > Async Processor Settings.
  3. Select DBI Failure Retry Hours.
  4. On the Property page, select the Node Type radio button and then in the column for your Landmark host, select the value (if one exists) or the <not set> link for the Async node.
  5. On the Edit Property form, enter a number for the value and click Create or Update. The value you enter is the number of hours for the retry limit.
  6. Save your changes.