RMI: can a client support cuncurrent calls from others client?

Question

Hi, I'm trying to build a simple peer 2 peer system.
My clients, when registering to the server, are passing a ResourceList (class that extends Remote) to the server, so that the server has a complete list of all the resources of the connected clients.
When a client search something, the server returns to him the list containing the searched resource(remote reference to the list of the owner client), and the searching client can call a getResource on the remote reference, that returns the Resource (serializable).
Now I'm wondering, if more than one client get the same remote reference to a list, and call getResource at the same time, what will happen?
Since each client for RMI is a thread, it's like n treads are calling the same method, that actually is remote.
So, if I synchronize this (remote) method, I can assure that only one client at time gets the Resource, but what if the client dies while having the lock on the resource?
Thanks in advance for any answer.

Answer 1

Now I'm wondering, if more than one client get the same remote reference to a list, and call getResource at the same time, what will happen? The same what happens in normal (non-remote) application with several threads calling the same method.
Since each client for RMI is a thread, it's like n treads are calling the same method. Right.
So, if I synchronize this (remote) method, I can assure that only one client at time gets the Resource,It depends what you mean by "gets the Resource". By synchronizing a method you assure that the execution of this method will be synchronized, but you cannot say this about the clients using the Resource later.
but what if the client dies while having the lock on the resource?Generally you don't have to worry about it, it will be handled by JVM. Either the client thread will die prematurely (and release the lock) or will execute normally and time out later when trying to send the response back to the client over (proably dead) connection. But in your case the resource will not be locked! You are synchronizing a method which basically assures that just a single thread may execute this method at a time.
Edited by: AdamDyga on Jun 24, 2010 3:59 PM