Failover unit

There is various clustering software available to cluster WASs and clusters. The way clustering solutions are implemented requires what we call a failover unit (there is no industry-acceptable term for this concept-each vendor uses a different name, such as package, resource group, service group, or virtual host). This failover unit is treated as a collection of computing resources that needs to run someplace. The operating system (and hardware) clustering solutions take responsibility for getting this collection to run on a system.

From a WebSphere perspective, the important concept is that the WebSphere processes run in a failover unit; they do not run on a machine. The actual machine where the WebSphere processes are running may vary over time, but the failover unit in which they run will not.

In general, a failover unit contains three things:
Applications
Access points
Dependencies

The list of applications for a failover unit contains the WebSphere processes (WAS process, WebSphere deployment management process, WebSphere node agent process, and WebSphere JMS server process). The access points to the applications are the entry points to the WebSphere processes. The dependencies are those things that are unique to a specific WebSphere process and must be moved when the process is moved. An example of such a dependency is the transaction log.

  Prev | Home | Next

 

WebSphere is a trademark of the IBM Corporation in the United States, other countries, or both.

 

IBM is a trademark of the IBM Corporation in the United States, other countries, or both.