# Free 300-620 practice questions

> 10 free 300-620 practice questions with the correct answer and an explanation.

Page: https://devopscodelab.com/exams/cisco-300-620/free-practice-questions/
Practice exams: https://devopscodelab.com/exams/cisco-300-620/

### Question 1 of 10

During initial fabric discovery, a newly cabled leaf appears in the APIC Fabric Membership view without a node ID. What must the administrator do so that the leaf receives its TEP address and joins the fabric?

- A. Configure an IP address on the leaf console and set its infra VLAN manually
- B. Assign a node ID and node name to the switch serial number to register it
- C. Create a static switch profile for the serial number under Fabric Access Policies
- D. Reload the leaf with the fabric discovery boot variable so it requests a new TEP

**Answer:** B

**Explanation:** Discovered switches stay unregistered until an administrator assigns a node ID and name (or preprovisions the serial number). After registration, the APIC's DHCP service hands out a TEP address from the pool and the switch downloads its policy. Console IP configuration and reloads are not part of normal ACI discovery, and switch profiles in access policies select already-registered node IDs for interface configuration, not registration.

### Question 2 of 10

An administrator must configure 20 server-facing ports on leaves 101 and 102 for a new rack. Which ordered chain of access policy objects connects a VLAN pool to the physical interfaces?

- A. VLAN pool, AAEP, physical domain, switch profile, interface policy group, interface profile
- B. VLAN pool, interface policy group, physical domain, AAEP, switch profile, interface profile
- C. VLAN pool, physical domain, AAEP, interface policy group, interface profile, switch profile
- D. Physical domain, VLAN pool, interface profile, AAEP, switch profile, interface policy group

**Answer:** C

**Explanation:** The VLAN pool is associated with a domain, the domain is linked to an Attachable Access Entity Profile, the AAEP is referenced by the interface policy group, the policy group is applied to port selectors in an interface profile, and the interface profile is associated with a switch profile that selects the leaf nodes. The other orders misplace the AAEP or the domain relationships and do not reflect how the objects reference one another.

### Question 3 of 10

A border leaf with an L3Out learns many remote endpoints from traffic arriving from other leaves, which causes stale entries after endpoint moves. Which global setting is recommended to address this?

- A. Disable Remote EP Learn so border leaves do not learn remote IPs from data traffic
- B. Enable rogue endpoint control with a lower detection interval across all leaves
- C. Increase the remote endpoint aging timer on every leaf of the fabric to 900 seconds
- D. Configure Flood in Encapsulation for every bridge domain associated with the L3Out

**Answer:** A

**Explanation:** Disabling Remote EP Learn (for border leaves) stops border leaves from learning remote IP endpoints from data plane traffic, avoiding stale entries; Cisco recommends it together with Enforce Subnet Check. Longer aging would keep stale entries longer, rogue endpoint control targets flapping endpoints, and Flood in Encapsulation scopes flooding rather than remote learning.

### Question 4 of 10

A bridge domain contains silent hosts that never send traffic until they receive ARP requests. ARP flooding is disabled and unicast routing is enabled. How does ACI reach a silent host whose IP is unknown?

- A. The spine glean process makes leaves with the BD subnet send ARP for the target IP
- B. The spine drops the ARP request, and the host stays unreachable until it sends traffic
- C. The leaf floods the original ARP request to all ports in the VRF, not only the BD
- D. The leaf converts the ARP request to unicast and sends it to the APIC for a response

**Answer:** A

**Explanation:** When the spine proxy has no entry for a target IP in a BD with unicast routing and a subnet, it triggers ARP glean: leaves with the BD generate ARP requests for the target from the BD gateway, which makes the silent host respond and be learned. The APIC is not in the data path, ARP is not flooded across the VRF, and silent hosts are reachable through glean.

### Question 5 of 10

An L3Out uses OSPF area 0.0.0.1 configured as NSSA. The engineer wants ACI to send only a default route into the area rather than all fabric routes. Which configuration achieves this design goal?

- A. Set the external EPG subnet 0.0.0.0/0 with Export Route Control for the area
- B. Set the BD subnets to private so that they are never advertised through any L3Out
- C. Configure the OSPF area type as totally stubby NSSA with no redistribution
- D. Configure the area as a regular area and filter routes with an import route map

**Answer:** A

**Explanation:** Route advertisement out of an L3Out is controlled by export route control; an external EPG subnet 0.0.0.0/0 with Export Route Control Subnet advertises a default route. Making BD subnets private stops advertising them but does not create a default route, an import route map filters inbound routes, and changing the area to totally stubby alters what ACI receives from the area.

### Question 6 of 10

Two L3Outs in the same VRF connect to different external networks. External EPG A on L3Out-1 has 0.0.0.0/0 for classification, and external EPG B on L3Out-2 also has 0.0.0.0/0. Traffic is misclassified. What is the recommended fix?

- A. Enable import route control on both L3Outs so classification follows the learned routes
- B. Move each L3Out to a separate border leaf, which makes the 0.0.0.0/0 classification local per leaf
- C. Add both external EPGs to the preferred group so classification no longer matters for policy
- D. Use specific prefixes per external EPG; 0.0.0.0/0 cannot classify for two L3Outs in one VRF

**Answer:** D

**Explanation:** External EPG classification is VRF-wide, by prefix, so the same prefix (such as 0.0.0.0/0) on two L3Outs in one VRF leads to ambiguous classification; specific prefixes should be used to separate them. Placing them on different leaves does not make classification per-leaf in general, preferred group avoids contracts but not misclassification, and import route control affects routes rather than classification.

### Question 7 of 10

An EPG associated with a VMware VMM domain uses Immediate resolution and Immediate deployment. A blade chassis with an intermediate switch that consumes LLDP sits between the ESXi hosts and the leaves. VMs in the EPG have no connectivity. What is the most appropriate fix?

- A. Change the VMM domain to a physical domain and bind the EPG statically to the blade uplinks
- B. Use Pre-provision resolution, or enable LLDP/CDP passthrough so hosts are discovered
- C. Change deployment immediacy to On-Demand so hardware is programmed when traffic arrives
- D. Increase the VLAN pool range so that the dynamic VLAN can be allocated to the port group

**Answer:** B

**Explanation:** Immediate resolution requires the APIC to correlate the host with the leaf port using LLDP or CDP; when an intermediate switch hides that, the policy is never pushed. Pre-provision (or a VMM domain with correct neighbor information) solves it. Changing deployment immediacy does not help when resolution fails, the VLAN pool is not the issue, and switching to a physical domain loses VMM integration.

### Question 8 of 10

A security policy requires that OOB management access to all switches be allowed only from 10.200.0.0/24 and only for SSH and HTTPS. An engineer creates an OOB contract with SSH and HTTPS filters, but access still works from 10.50.0.0/24. What is the likely cause?

- A. The OOB EPG consumes the contract instead of providing it, which reverses the filtering direction
- B. The contract scope is set to VRF instead of Global, which allows all subnets
- C. The external management instance profile still includes a broad subnet like 0.0.0.0/0
- D. The SSH filter uses TCP source port 22, which matches all management traffic

**Answer:** C

**Explanation:** Which sources may use the OOB contract is defined by the subnets in the external management network instance profile; if it still contains 0.0.0.0/0 or a broader range, other subnets keep access. Scope does not open other subnets, a wrong provider/consumer relationship would block rather than permit, and a source-port filter would not broaden access to other subnets.

### Question 9 of 10

An administrator wants to ensure that if all TACACS+ servers are unreachable, administrators can still log in with local accounts. Which feature provides this?

- A. The fallback login domain, which uses local authentication
- B. SNMPv3 users with authPriv credentials on each controller
- C. A second TACACS+ provider added to the same remote login domain
- D. Configuring the default authentication realm as local

**Answer:** A

**Explanation:** APIC provides a 'fallback' login domain that uses local authentication, so administrators can log in as apic:fallback\\admin if remote servers fail (when fallback check settings allow). Setting the default realm to local changes the default for all users, adding a second provider does not help if all are unreachable, and SNMP users cannot log in to the GUI.

### Question 10 of 10

An NDO template containing VRF Prod and BD Web is associated with Site 1 and Site 2. What is the result?

- A. Objects are created in both sites but are not linked to each other
- B. VRF Prod and BD Web are stretched and created in both sites
- C. BD Web is created only in the site with the most endpoints
- D. VRF Prod is created only in the first site in the template

**Answer:** B

**Explanation:** Objects in a template associated with multiple sites are deployed to each site and treated as stretched objects, with NDO handling the translation between sites. They are not limited to one site, not placed based on endpoint counts, and they are linked through namespace mapping.

Ready for a timed run? Take the free 15-question 300-620 mini exam: https://devopscodelab.com/app/exams/cisco-300-620/mini/

DevOpsCode Lab is not affiliated with Cisco. 300-620 and related names are trademarks of their owners. We write every question ourselves; these are not real exam questions.
