Automated provisioning of LAGs between AHV and VSX is an operation supported when Aruba Fabric Composer (AFC) is integrated with Nutanix. AFC is a tool that provides automation and orchestration for managing data center networks composed of ArubaOS-CX switches. AFC can integrate with various data center software such as VMware vSphere, Nutanix AHV, Microsoft Hyper-V, etc. AFC can discover, monitor, and configure Nutanix AHV clusters and hosts using REST APIs. AFC can also automate the provisioning of LAGs between AHV and VSX by creating VSX LAGs or MC-LAGs on the ArubaOS-CX switches and configuring the corresponding LAGs on the AHV hosts1.

Place all three VRFs in the same OSPF process on Switch-1 is not a valid strategy for meeting this goal of enabling devices in VRF B and VRF C to reach shared resources in VRF A. This strategy would not work because OSPF does not support multiple VRFs in the same process on ArubaOS-CX switches. Each VRF must have its own OSPF process with a unique process ID1.

Switch-2 waits a period called the link-up delay before it enables Switched Virtual Interfaces (SVIs) on its VSX LAGs is a true statement about what happens when Switch-2 experiences a power failure while Switch-1 remains up and then recovers. Switch-1 and Switch-2 are ArubaOS- CX switches, which are part of a Virtual Switching Extension (VSX) fabric. VSX is a feature that provides active-active forwarding and redundancy for ArubaOS-CX switches. The link-up delay timer defines how long a VSX node waits before advertising link state changes to its peer node.
This allows the node to synchronize its MAC forwarding, ARP, and routing tables with its peer node before sending or receiving traffic on the newly activated link.

VXLAN is a tunneling protocol that encapsulates layer 2 traffic over an IP network using VXLAN Network Identifiers (VNIs) to identify different layer 2 segments. VXLAN Tunnel Endpoints (VTEPs) are devices that perform the encapsulation and decapsulation of VXLAN packets. VTEPs can use different methods to handle broadcast, unknown unicast, and multicast (BUM) traffic within a VNI. One of these methods is headend replication, which means that the VTEP that receives a BUM packet replicates it and sends it as a unicast to each remote VTEP in the same VNI1. This method does not require multicast routing in the underlay network, but it can increase the load on the ingress VTEP. Therefore, this correctly describes how the VTEPs handle VXLAN traffic forwarding without EVPN.