Many tests for pings or connectivity are in a loop that exits when the first failure occurs. The loop should continue and then report a final failure if needed. This allows better debugging as many times the failures are on a single node. Seeing certain ips for vms on these nodes is helpful to identify the single node.