Heuristics during recovery used to be treated via the OTS hooking in various notification mechanisms, e.g., the CORBA notification service. That code is gone and as a result we silently ignore heuristics during recovery. This is not an issue for consistency, because the heuristic resolution tool(s) were always intended to scan the object store and flag all logs that are in a heuristic state. However, it might be nice for admins to have a pro-active indication that a failure has occurred. Need to check all recovery instances for those that are now silent.