Application resilience is one of the key elements
in a clustered environment. If you are planning to write and use highly
available applications in your cluster you should be aware that these applications
have specific availability specifications. 
By taking advantage of resilient applications in your cluster, an application can be restarted on a different cluster node without requiring you to reconfigure the clients. In addition, the data that is associated with the application will be available after switchover or failover. This means that the application user can experience minimal, or even seamless, interruption while the application and its data switch from the primary node to the backup node. The user does not need to know that the application and data have moved on the back end.
In order to achieve application resiliency
in your cluster, applications that meet certain availability specifications
must be used. Certain characteristics must be present in the application in
order for it to be switchable, and therefore always available to the users
of the application in the cluster. See High Availability and Clusters for details on these
application traits. Because these requirements exist, you have the following
options for using a switchable application in your cluster:
Once you have a resilient application, it must be managed within your cluster.