# Free HPE6-A86 practice questions

> 10 free HPE6-A86 practice questions with the correct answer and an explanation.

Page: https://devopscodelab.com/exams/hpe6-a86/free-practice-questions/
Practice exams: https://devopscodelab.com/exams/hpe6-a86/

### Question 1 of 10

Host A is supposed to be configured as 10.1.1.10/24 with default gateway 10.1.1.1, but a typo set its mask to 255.255.0.0. Host A now tries to reach server 10.1.2.20, which is in a different VLAN behind the router, and proxy ARP is disabled. What happens?

- A. Host A treats 10.1.2.20 as local, sends an ARP request for it directly, and gets no reply
- B. Host A broadcasts the IP packet itself to every host in 10.1.0.0/16 until the server replies
- C. Host A sends the packet to its default gateway, which routes it normally to the server
- D. Host A sends an ARP request for 10.1.1.1 and the router replies with the server's MAC

**Answer:** A

**Explanation:** With a /16 mask, Host A believes 10.1.2.20 is on its own subnet, so it ARPs for the server directly instead of for the gateway; because the server is in another VLAN and proxy ARP is off, no reply arrives and the traffic fails. The host never uses its gateway for destinations it considers local, a router answers ARP only for its own address unless proxy ARP is enabled, and hosts do not broadcast unicast IP packets.

### Question 2 of 10

Users report that their PCs receive addresses from an unexpected 192.168.0.0/24 range after a visitor plugged in a consumer router. Which switch feature prevents this from recurring?

- A. DHCP snooping with trusted ports only toward the legitimate DHCP servers
- B. LLDP-MED network policies that push the correct VLAN to every connected PC
- C. Dynamic ARP inspection configured on every access port without DHCP snooping
- D. An IP helper address on each user VLAN pointing to the corporate DHCP server

**Answer:** A

**Explanation:** DHCP snooping drops DHCP server messages arriving on untrusted ports, so only the ports facing the real DHCP servers can offer leases. ARP inspection relies on the DHCP snooping binding table and does not block rogue offers by itself, an IP helper forwards requests to the right server but does not stop a rogue server from answering, and LLDP-MED policies apply to phones rather than DHCP.

### Question 3 of 10

An architect must choose between a VSF stack of CX 6300 switches and a VSX pair of CX 6400 chassis for a building's aggregation layer. The requirement is that a software upgrade must never take the whole aggregation layer offline at once. Which choice best meets this and why?

- A. The VSX pair, because both peers merge into one control plane while each one is upgraded
- B. The VSF stack, because all members share one control plane and upgrade simultaneously
- C. The VSF stack, because the standby member keeps forwarding while the conductor upgrades
- D. The VSX pair, because each peer has its own control plane and can be upgraded separately

**Answer:** D

**Explanation:** VSX keeps independent control planes, so one peer can be upgraded and rebooted while the other continues forwarding, which avoids a full-layer outage. A VSF stack shares one control plane, so a stack-wide upgrade affects all members; claiming that VSF upgrades members while forwarding, or that VSX merges control planes, misstates how the technologies work.

### Question 4 of 10

A network operations team wants a CX switch to watch for a condition such as rising CPU use on a routing process, keep a history of the metric, and automatically collect diagnostic output when the threshold is crossed. Which built-in AOS-CX feature provides this?

- A. Network Analytics Engine (NAE) agents
- B. Checkpoints with an automatic rollback timer
- C. sFlow sampling sent to an external collector
- D. Port mirroring to a remote packet analyzer

**Answer:** A

**Explanation:** NAE runs agents on the switch that monitor database values, store time-series history, and trigger actions such as collecting show output or sending alerts when conditions are met. Checkpoints save and restore configurations, port mirroring copies traffic, and sFlow exports sampled traffic to a collector without on-switch analytics or automated actions.

### Question 5 of 10

IP phones on CX access ports should learn their voice VLAN, 100, automatically through LLDP-MED, while PCs behind them stay in data VLAN 20. Which configuration approach fits AOS-CX?

- A. Configure VLAN 100 as the native VLAN on each port and leave VLAN 20 untagged as well
- B. Enable dhcpv4-snooping on VLAN 100 so DHCP options assign phones to the voice VLAN
- C. Mark VLAN 100 with the voice keyword and allow it tagged on the ports with VLAN 20 as native
- D. Create an SVI on VLAN 100 so LLDP-MED advertises its IP address as the voice VLAN ID

**Answer:** C

**Explanation:** Defining VLAN 100 as a voice VLAN lets AOS-CX advertise it through LLDP-MED network policy, and the port carries VLAN 20 untagged for the PC and VLAN 100 tagged for the phone. A port cannot have two untagged VLANs in this way, an SVI address is not the voice VLAN advertisement, and DHCP snooping does not assign VLANs.

### Question 6 of 10

Two aggregation switches run MSTP. VLANs 10-19 should prefer AGG1 as root and VLANs 20-29 should prefer AGG2, with the other switch as backup in each case. Region settings already match on all switches. Which configuration on AGG1 supports this?

- A. spanning-tree instance 1 vlan 10-29 with instance 1 priority 0 only, leaving instance 2 unchanged on AGG1
- B. spanning-tree instance 1 vlan 10-19 with instance 1 priority 15, and instance 2 vlan 20-29 with instance 2 priority 0
- C. spanning-tree priority 0 for the CIST only, because all MST instances inherit the CIST root
- D. spanning-tree instance 1 vlan 10-19 with instance 1 priority 0, and instance 2 vlan 20-29 with instance 2 priority 1

**Answer:** D

**Explanation:** On AGG1, mapping each VLAN range to its own instance and giving instance 1 the lowest priority (0) while instance 2 gets a slightly higher value (1, with AGG2 configured as 0) makes AGG1 primary root for VLANs 10-19 and backup for 20-29. Mapping all VLANs to one instance prevents load sharing, the CIST priority does not set roots for individual MST instances, and priority 15 is the highest multiplier, which makes AGG1 least preferred for instance 1.

### Question 7 of 10

802.1X authentication fails for every client on a newly deployed CX switch. The RADIUS server log shows requests arriving from the switch but rejected as having an invalid message authenticator. What is the most likely cause?

- A. Client supplicant certificates have expired
- B. The RADIUS server's UDP port 1813 is blocked by a firewall in the path
- C. The switch's shared secret does not match the RADIUS server's
- D. The switch is sending requests from the mgmt VRF instead of the default

**Answer:** C

**Explanation:** An invalid message authenticator means the RADIUS server cannot validate the request with its configured secret, which indicates a shared secret mismatch with the switch. Expired client certificates would fail individual users with a different reason, the requests already arrive so the source VRF is not the issue, and port 1813 is accounting rather than authentication.

### Question 8 of 10

After a power outage, a CX VSF stack of three members came back as two separate devices: member 1 alone and members 2 and 3 together. Split detection was configured. Which outcome indicates split detection worked correctly?

- A. Both fragments reboot repeatedly until the VSF links are recabled by a technician
- B. Both fragments keep forwarding with the same IP addresses to keep users connected
- C. Only one fragment keeps its interfaces active, and the other disables its ports
- D. The fragment with the higher member number becomes active and the other one resets

**Answer:** C

**Explanation:** VSF split detection keeps one fragment active and makes the other shut down its network interfaces, preventing duplicate IP addresses and conflicting forwarding. Two active fragments with the same addresses is exactly the problem it prevents, endless reboots are not its behavior, and higher member numbers are not the deciding factor.

### Question 9 of 10

A help desk group must be allowed to bounce access ports and view status on CX switches, but must not change VLANs, routing, or AAA settings. The built-in operators group is too limited and administrators is too broad. What should be configured?

- A. The auditors group, which allows interface changes but blocks routing changes
- B. A custom user group with command rules that permit only the required commands
- C. The administrators group for help desk staff, with a policy asking them to be careful
- D. A separate VRF for the help desk so their commands only affect access ports

**Answer:** B

**Explanation:** AOS-CX supports user-defined groups whose command rules allow or deny specific CLI commands, matching the required privileges precisely. Giving administrator rights violates least privilege, auditors are meant for viewing logs rather than changing interfaces, and VRFs separate routing tables rather than user privileges.

### Question 10 of 10

An administrator wants the CX switch to raise an alert when a specific uplink's utilization stays above 80 percent for several minutes, using only on-switch capabilities. Which feature fits?

- A. A mirror session that copies the uplink traffic to a collector
- B. A checkpoint that records the uplink's configuration each hour
- C. An NAE agent that monitors the interface and triggers an action
- D. An LLDP neighbor entry that lists the uplink's partner switch

**Answer:** C

**Explanation:** NAE agents can monitor interface statistics over time and trigger alerts or actions when a condition such as sustained high utilization is met. Mirroring copies traffic but needs external analysis, LLDP only identifies the neighbor, and checkpoints capture configuration rather than utilization.

Ready for a timed run? Take the free 15-question HPE6-A86 mini exam: https://devopscodelab.com/app/exams/hpe6-a86/mini/

DevOpsCode Lab is not affiliated with HPE. HPE6-A86 and related names are trademarks of their owners. We write every question ourselves; these are not real exam questions.
