Posted By:
nuno_oliveira
Posted On:
Friday, January 3, 2003 12:50 PM
Consider the following setup: - two hosts A and B. - start the rmiregistry in host B in a command prompt. - in a diferent command prompt in this same host B, start a program that registers a remote server implementation reference in the above registry. - in host A, start a client that contacts the rmiregistry running in B to get a reference to the server object and waits for user input before exiting. - kill the VM in host B by closing the console window where you started the server (or using ctrl-c, whatever...). My understanding of RMI's distributed garbage colection tells me that if you leave the client on for enough time, RMI's automatic lease renewal should detect that the RMI-runtime on host B i
More>>
Consider the following setup:
- two hosts A and B.
- start the rmiregistry in host B in a command prompt.
- in a diferent command prompt in this same host B, start a program that registers a remote server implementation reference in the above registry.
- in host A, start a client that contacts the rmiregistry running in B to get a reference to the server object and waits for user input before exiting.
- kill the VM in host B by closing the console window where you started the server (or using ctrl-c, whatever...).
My understanding of RMI's distributed garbage colection tells me that if you leave the client on for enough time, RMI's automatic lease renewal should detect that the RMI-runtime on host B is down (since the VM has terminated) and throw an exception on the client side.
Is this true ? My tests say it isn't... maybe it's only a question of waiting a litle while ?
Thank you.
<<Less
