After an ESXi host is rebooted, virtual machines across multiple VLANs lose network connectivity. The affected VMs can ping other VMs on the same subnet, including VMs running on other hosts, but cannot ping their default gateway, which is another virtual machine.
Reviewing ARP on an affected VM shows that the VM resolves ARP for other VMs on the same subnet, but the default gateway does not appear in the ARP table. The gateway in these environments is a firewall running as a virtual machine, and its virtual network adapters came up disconnected after the host reboot, so it no longer responds to ARP.
Because that firewall VM is the default gateway for VMs across multiple VLANs, every VM that routes through it loses connectivity at the same time, even though those VMs are otherwise healthy. The result presents as a broad, multi-VLAN outage rather than a single-VM problem, which can make the gateway VM easy to overlook as the source.
One or more virtual network adapters on the gateway virtual machine come up in a disconnected state after the host reboot. With the gateway VM offline at the network layer, the VMs that depend on it for their default gateway cannot reach it and lose connectivity.
Reconnect the virtual network adapters on the affected gateway virtual machine:
Verifying Connect at power on prevents the adapters from coming up disconnected again after a subsequent host reboot or power cycle.
If the issue persists after following these steps, contact Broadcom Support for further assistance.