|
|
The Node Agents and the Deployment Manager are administrative processes in the WebSphere environment. One of their main responsibilities is to keep configuration data synchronized. Node Agents and the Deployment Manager both use a repository of XML files on their own nodes. The master repository data is stored on the Deployment Manager node. That data is then replicated to each node in the cell. The default synchronization interval is 60 seconds. The synchronization process is unidirectional from the Deployment Manager to the Node Agents to ensure repository integrity. That means that any changes made on the Node Agent level are only temporary and will be overwritten by the Deployment Manager during the next synchronization. Only changes to the master data through the Deployment Manager are permanent and replicated to each node.
Other tasks for the Deployment Manager are related to the Tivoli Performance Viewer, backup clusters, JMX routing, distributed logging, the naming server and the security server. To learn what impact a failure of the Deployment Manager has on the various WebSphere system components, see 3.2, Deployment Manager failures.
In addition to the file synchronization services, the Node Agent also provides runtime support for the Location Service Daemon (LSD), JMX server, distributed logging, naming server, security server, and is required for starting appservers. See 3.3, Node Agent failures for details.
You learn in this chapter that, depending on your requirements and WebSphere environment, the impact of a failure of one of these processes might be minimal. Some customers even run their production environment without an active Deployment Manager.
