The administrator cannot ping the hostname of any system on the local network, including LPARs in the same managed system that have virtual Ethernet adapters configured on the same subnet and VLAN…

A disconnected or faulty physical Ethernet cable would also break IP-level connectivity, but LPARs sharing a virtual Ethernet adapter on the same VLAN within the same managed system communicate through the hypervisor's internal virtual switch without traversing the physical cable.

A Shared Ethernet Adapter failure in the VIO Server would disrupt traffic between the internal virtual network and the external physical network, but intra-VLAN communication between LPARs on the same managed system does not traverse the SEA, so those LPARs would still be reachable by IP.

A flushed routing table removes default and static routes needed to reach other subnets, but systems on the same local subnet are reached via direct layer-2 forwarding and do not require a routing table entry, so same-subnet LPARs would remain pingable by IP.

Because LPARs on the same VLAN and subnet would still be reachable by IP through the virtual switch regardless of physical cabling or routing issues, the fact that hostname-based pings fail for all hosts including those on the same subnet isolates the problem to name resolution - when DNS is disabled and /etc/hosts lacks the required hostname-to-address mappings, the OS cannot resolve any hostname before attempting the ping.