I have a master running and a remote slave connected to it. Then the connection between the master and slave is broken. If the connection is then re-established, the slave will attempt to connect again and fail, because the master says the slave is already connected. A message is shown saying:

Exception: The server rejected this connection: <machine> is already connected to this master. Rejecting this connection at hudson.remoting.Engine.run(Engine:191)

Then if a new job run is started, the original connection to the slave fails saying "connection interrupted" and there is then no connections with the slave.