Archived

This forum has been archived. Please start a new discussion on GitHub.

Beginner question

Hi,
I created an ICE callback server through glacier as firewall. My silverlight client connects to it and waits for call back from the server.

Everything is ok. Server runs without any issues and client is able to receive updates. I also do a session.ice_ping every 50 seconds to make sure that the client object is not disposed.

I have a couple of questions beyond this.
1. If the ICE server goes down and glacier is running, is there any way for the client to know that the server is down? Currently I could not figure it out.

2. Also, when I create a server object and wait for clients to connect will every client connection get one instance each of the ICE object I created or will all the clients connect to the same instance of Ice remote object?

Comments

  • benoit
    benoit Rennes, France
    Hi,
    muthu.v wrote: »
    1. If the ICE server goes down and glacier is running, is there any way for the client to know that the server is down? Currently I could not figure it out.

    Yes, the client session.ice_ping() call will fail with an exception when the server becomes unreachable.
    2. Also, when I create a server object and wait for clients to connect will every client connection get one instance each of the ICE object I created or will all the clients connect to the same instance of Ice remote object?

    Assuming you're not using servant locators (but only servants registered with the object adapter), all the client invocations for the Ie object will end up being dispatched by the same servant instance: the one you have registered with one of the object adapter add methods.

    Cheers,
    Benoit.