|
|
The first option to enhance the Node Agent's availability is to monitor the process using the operating system' capabilities. See 3.4, Restarting WebSphere processes as an OS service.
The configuration steps to make WAS highly available using a platform-specific clustering technique vary from product to product. But you always have to define some kind of failover unit. These units are switched between the nodes in case of a failover. Usually the failover units need a start, stop, and monitor program. You find detailed instructions on how to set up HACMP, Sun Cluster, Tivoli System Automation, and VERITAS Cluster Server in the appropriate chapters of this part.
