At the foundation of continuous operation is the no-single-point-of-failure (SPOF) rule. That is, no single hardware or software component failure exists that can appear to the end user as an application outage.

That does not mean that the user does not see a failure or have to connect again, but that the user can quickly resume their previous work with minimal interruption. What you are looking for are architectural components that appear only once in your application’s deployment. A good place to start is looking for and understanding those SPOFs in:
  • File system data storage
  • Networking components
  • OS image
    • Database server
    • Application servers

The types of SPOF that you can incorporate into your ABL application’s architecture are the OS images and the OpenEdge products that operate inside one. Following the basic SPOF rule, you should have a plan to have two, or more, instances of each OpenEdge component running and able to take over the load in case of failure.

Most OS images have failover or recovery for simple daemon processes, whether it is simple built-in service process monitoring or more container monitoring when your daemon services are running in that environment. Look at your OS image’s feature and package support for the best fit for your ABL application’s processes.