daemon loses request during restart
If a guardian daemon is completely restarted, it looses whatever its request state was before the restart.
Typically managers only set the request once during main(), and don't continually set the request during run(). The problem with continually resetting the request is that there is a potential race condition with a managed node that resets it's request after a jump. If the manager continually sets the request it might reset the request after jump, which might cause the subordinate to get back to the request before the manager had a chance to ever notice that a jump transition occurred.
So, it would be good if there was another way to reacquire the last request after restart.