Hi Charan,
Let us try to get the questions one by one.
are you asking from EJB stub perspective or servlet perspective?
In the case of servlets, you would have configured proxy_server which is aware of all the nodes in the cluster. So the incoming request would have the proxy_server address. Proxy based on the heartbeat mechanism will route it to appropriate managed node. If the managed node is down, it will communicate to the next one. Various algorithms are there for doing this like round-robin, weighted round robin, random, etc.
In the case of EJB, at the time of compilation itself the server can determine what are the cluster nodes in the cluster and incorporate them into stubs. So when you connect to this EJB, the stubs themselves know what are the nodes in the cluster and if one goes down, it will automatically contact the second and so on.
SerializedSystemIni.DAT - This is more of internal details of weblogic which user need not bother about unless otherwise he has very specific reason. Are there any specific question you have on this?
