This issue may occur when a physical switch in the network fabric stops processing traffic but maintains an active physical port link state (link-up).
When an ESXi host is configured with a static port channel ("Route based on IP Hash"), it relies entirely on the physical link status for uplink failure detection. Because static port channels lack a dynamic health-checking mechanism, the ESXi host is unaware of the upstream traffic flow failure. The host's load balancing algorithm continues to forward traffic for specific IP hashes to the unresponsive switch, creating a traffic blackhole.
Investigate with the network team to identify and resolve the underlying physical network issue.
Long-Term Prevention: To prevent silent hardware failures from causing network outages, migrate the ESXi networking configuration from Static Port Channels to Link Aggregation Control Protocol (LACP).
Implement a vSphere Distributed Switch (vDS), which is required for LACP support in vSphere.
Configure Link Aggregation Groups (LAG) with LACP on the vDS and the physical switch fabric. LACP utilizes Protocol Data Units (PDUs) to actively negotiate and monitor link health, automatically removing unresponsive uplinks from the active forwarding path even if the physical link remains electrically active.
Understanding IP Hash load balancing (KB 321396)
Host requirements for link aggregation (etherchannel, port channel, or LACP) in ESXi (KB 324555)
Enable EtherChannel / Link Aggregation Control Protocol (LACP) in vSphere (KB 321425)
Network Loss on ESXi During Physical Switch Reboot or Hardware Failure Without Link State Change