300-610 — Frequently Asked Questions

Twinax (DAC) cables are typically used for server uplink ports (L3/L4). The dedicated cluster ports (L1/L2) are designed for standard Ethernet SFP+ modules and cabling.

Cross-connecting L1 to L2 ensures that if one internal switch plane fails, the other path remains active, providing true redundancy for the cluster control plane.

FCoE frames require specific tagging and processing that conflicts with native VLAN handling, which expects untagged frames. Cisco NX-OS prohibits FCoE on native VLANs.

No. While it avoids the native VLAN restriction, you must still map the VLAN to a VSAN in the switch's VSAN database for FCoE to function.

Uplink pinning ensures consistent path selection for specific traffic flows, simplifying troubleshooting and ensuring predictable failover behavior in iSCSI environments.

Yes, if you already own them, they are valid. However, the question specifies 'not require purchase of new equipment,' implying we must use what is available or software alternatives.

Switchport mode E creates an Edge port, which terminates traffic and does not support trunking multiple VSANs across a single physical link.

FC trace is typically utilized on trunking links (TE) where multiple VSANs are active, allowing for better visibility into the fabric traffic flow.

While MTU 9000 improves throughput, the question prioritizes minimizing oversubscription and latency. Port binding (B) handles load balancing/oversubscription, and fewer hops (E) directly minimizes latency.

Yes, software iSCSI relies on the host CPU for packet processing, which increases latency and CPU load compared to hardware-assisted or offloaded iSCSI implementations.

RDMA bypasses the CPU stack entirely, whereas the question specifically asks for improvements to the CPU's own data cache hit rate via steering.

ARFS provides hardware-assisted steering that updates the kernel's flow table dynamically, ensuring packets are processed on the core where the socket is open, improving cache locality.

Yes. Thick provisioning allocates space upfront but does not prevent snapshot creation. HyperFlex manages snapshots at the block level regardless of initial provisioning.

Duplicate names within the same vCenter scope are generally not supported for deployment, but the specific documented blocker for snapshot failures is the suspended state.

vPC requires manual peer-link and peer-switch configuration, while vPC+ automates these aspects when connecting to an ACI or VXLAN fabric.

No, standard vPC is not optimized for direct integration with VXLAN fabrics without significant complexity; vPC+ is the intended feature.

VXLAN provides the data plane encapsulation but lacks the native control plane features for efficient MAC route distribution and loop prevention that EVPN offers.

OTV is a Cisco proprietary technology, which conflicts with the requirement to use equipment from different manufacturers to prevent vendor lock-in.

An FCoE uplink port connects the fabric interconnect northbound to an upstream FCoE switch or SAN fabric, not directly to a storage array; using it would imply the array sits behind that upstream device.

No. Fibre Channel storage ports are for native FC speeds (8/16/32 Gbps) and cannot carry FCoE or the lossless class-fcoe Ethernet QoS system class required here.

Direct-attach storage does not support SAN port-channeling. Using fewer ports (two) ensures proper redundancy across FIs without relying on unsupported FC switch features.

The 32G FC ports on the Fabric Interconnects connect directly to the storage array controllers, handling the actual storage traffic, separate from the IOM uplinks.

That command is used under BGP neighbor configurations to advertise the EVPN address family, not to define route targets within a specific VRF instance.

It automatically generates a Route Distinguisher and Route Target based on the VRF name, applying them for both import and export of EVPN routes.

FCoE encapsulates native Fibre Channel frames. Since it remains part of the FC fabric, standard FC zoning rules apply to define access permissions.

No. A Storage port configures the FI as an FC switch. Zoning is a fundamental mechanism of FC fabrics and cannot be disabled while using this port type.

End-host mode is optimized for connecting to upstream fabrics, not direct storage. Direct storage attachment typically requires the FI to act as a switch (Switching mode) to manage the fabric services.

Storage ports connect directly to storage arrays in switching mode. Uplink ports connect to upstream Fibre Channel fabrics, often in end-host mode.

GLBP is Cisco-proprietary, so it cannot run on the customer's routers from a variety of vendors. VRRP is the open-standard option that still supports tunable load sharing via multiple groups.

FHRP is only the generic family name for VRRP, HSRP and GLBP, not a protocol you configure. The question asks which solution to deploy, so the specific standard-based protocol is required.

BGP can leak routes, but it is a Layer 3 mechanism. VXLAN is an overlay that extends Layer 2, making it more suitable for presenting shared services (which often rely on broadcast/multicast) to multiple isolated segments.

No, but in the context of modern Data Center designs (PCSE/PCNSE), VXLAN EVPN is the standard method for stretching L2 domains and sharing services across VRFs without complex routing policies.

The question asks for VNI-to-multicast mappings, not NVE mappings. NVEs are the endpoints, while VNIs are the logical networks being mapped to groups.

It significantly reduces the number of multicast groups needed in the underlay network, improving scalability and reducing control plane overhead.

VNIC placement pins vNICs to specific hardware paths but doesn't split interrupt handling across cores unless RSS is also enabled.

Most modern VIC adapters support RSS, but it must be explicitly enabled in the Ethernet adapter policy within UCS Manager.

FCoE is a transport protocol. While it enables connectivity, RDM is the specific virtualization technology used to map the external LUNs to the VMs for clustering awareness.

While technically possible in some setups, the question specifies an external SAN fiber channel array, making RDM via the FC path the intended architectural solution.