Hard 350-601 practice questions

In a Cisco ACI fabric, an application team reports that a legacy clustering application relying on gratuitous ARP and silent hosts is failing intermittently after workloads were migrated into a new bridge domain (BD). The BD was created with the default optimized settings. Which BD configuration change should the network engineer make to best support the silent hosts and the application's ARP-based behavior?

An ACI administrator is deploying two new EPGs, each mapped to a different physical domain. EPG-Web uses a physical domain tied to VLAN Pool-A (dynamic allocation, blocks 100-200), and EPG-App uses a physical domain tied to VLAN Pool-B (static allocation, blocks 150-250). After the administrator statically binds EPG-App to a leaf port using VLAN 175, EPG-Web endpoints intermittently lose connectivity. What is the most likely root cause?

A data center team is building a dedicated back-end fabric to interconnect 256 GPUs across multiple servers for large-model training. The AI workload is dominated by all-reduce collective operations that generate synchronized, bursty east-west traffic where the slowest flow dictates job completion time. Which fabric design characteristic is MOST important to prioritize to minimize tail latency and avoid degrading GPU utilization?

A data center runs VXLAN BGP EVPN across two leaf switches (Leaf1 and Leaf2) with ARP suppression enabled on the bridge domain's associated VLAN. Host A on Leaf1 attempts to reach Host B, a silent host connected to Leaf2 that has not yet sent any traffic. Host A sends an ARP request for Host B. Which behavior occurs?

A data center engineer is troubleshooting an ACI fabric where an endpoint in EPG-Web on Leaf-101 sends an ARP request for an endpoint in the same bridge domain that is currently unlearned by the fabric. The bridge domain is configured with ARP Flooding disabled and L2 Unknown Unicast set to Hardware Proxy. How does the ACI fabric handle this ARP request?

A storage administrator is extending a Fibre Channel ISL between two data centers 80 km apart over dark fiber at 8 Gbps. After bringing up the link, throughput is far below line rate even though no errors are logged and the link stays up. Which action is most likely to resolve the performance problem?

A storage administrator is extending a Fibre Channel SAN between two data centers 600 km apart using an FCIP tunnel over an existing IP WAN link. Users report that remote tape backup jobs are extremely slow, and analysis shows that each SCSI write command must wait for a status response before the next command is sent, causing the round-trip WAN latency to dominate throughput. Which FCIP feature should the administrator enable to improve backup performance across the tunnel?

A storage administrator is troubleshooting suboptimal path selection in a Fibre Channel fabric. Traffic between two MDS switches is traversing a 8 Gbps ISL instead of a parallel 32 Gbps ISL. Both ISLs are in the same VSAN and are up. What is the most likely cause, and what mechanism determines the preferred FC route?

A storage engineer is integrating a Cisco MDS 9000 fabric with a legacy third-party FC switch fabric that requires standards-based interoperability. The engineer must place the connecting VSAN into interop mode to allow the fabrics to merge without proprietary features breaking the ISL. What is a primary consequence the engineer must plan for when enabling interop mode 1 on the VSAN?

A data center engineer bundles four interfaces into a port-channel between a Nexus 9000 and a downstream server using LACP. The requirement is that if fewer than three member links are operational, the entire port-channel must go down rather than forward traffic on the remaining links. Additionally, if the server sends no LACP PDUs on a member interface, that interface must NOT independently forward traffic as a standalone access port. Which combination of configuration achieves both requirements?

A data center engineer notices that after a vPC domain reload on the secondary Nexus 9000 switch, all vPC member ports briefly went down and traffic was disrupted even though the type-1 consistency parameters matched on both peers. The engineer wants to prevent vPC member ports from being suspended on the recovering peer before it has fully synchronized state with the primary peer following a reload. Which vPC feature should be enabled to address this?

A network engineer wants an NX-OS device to automatically collect a detailed diagnostic bundle whenever an interface flaps more than five times in 60 seconds. The collection logic is complex, requiring conditional parsing of multiple show commands and formatting output into a JSON file stored in bootflash. The engineer plans to use an EEM policy to trigger this. Which approach best satisfies the requirement while keeping the trigger logic maintainable?

A data center team is deploying a GPU cluster for AI/ML training workloads across Nexus 9000 switches using RoCEv2 for RDMA transport. During large distributed training jobs, they observe throughput collapse and retransmissions caused by packet drops in the fabric. The team wants to build a lossless Ethernet fabric that also reacts to congestion before buffers overflow. Which combination of technologies should they implement to meet these requirements?

In a Nexus 9000 vPC domain, a switch receives a broadcast frame on a vPC member port connected to a downstream switch. The peer-link is up and the vPC is fully operational. How does the vPC peer that receives the broadcast forward it, and what rule governs delivery to the far-side vPC member ports?