Since netobjd is often run in background, it is easy to forget that an instance is running. The second instance of netobjd will fail with a TCP error when it attempts to open its listener socket.
If an instance of netobjd dies, all of the objects that were registered in its table are lost. There currently is no standard mechanism for restarting netobj after a failure. Even if netobjd is restarted, there is currently no stable record of the objects held by previous instantiations.