* enos: default undo-logs to cluster behavior * change a step dependency * rearrange steps, wait a bit longer for undo logs