Hello, Suppose I run snow on cluster using MPI as the backend. Does it handle the case of a compute node(the entire machine) failing and reassigning the task to the another machine? Regards Saptarshi