Same here. A single node failure has lead to the whole cluster crashing down around me on more than one occasion.