CCIE Enterprise Infrastructure practice lab 4 of 8
DMVPN Phase 3 dual-hub fault hunt
You inherit a dual-hub DMVPN with front-door VRFs, IKEv2 and EIGRP that has never worked properly. Find and fix nine faults, layer by layer, prove shortcuts and hub failover with a fault injector, then add a static GRE tunnel and an IPv6 overlay. A 3 to 4 hour practice lab that starts broken.
Run this lab
It needs Standard: its 9 nodes are set to 9 vCPU and 4,608 MB in total.
In your workspace: Shared, then CCIE Enterprise Infrastructure practice guided labs, then 04 - DMVPN Phase 3 dual-hub fault hunt
- Level
- Advanced
- Time
- About 210 minutes
- Nodes need
- 9 vCPU, 4,608 MB
- Smallest plan
- Standard
The scenario
Halden Freight's last network engineer left halfway through a WAN migration. The design below is what the company paid for. What's on the routers is what he left. Some sites have never joined, some work only through one hub, and nobody has ever tested a hub failure.
10.4.100.0/24
DC1 -------------- ISP (for the GRE tunnel only)
/ \
HUB1 HUB2 primary and backup hub
\ /
+---------+
| ISP | Internet, public addresses only
+---------+
/ | | \ \
SP1 SP2 SP3 SP4 PARTNER| Design item | Standard |
|---|---|
| Underlay | Each hub and spoke WAN interface sits in VRF INET with a default route to ISP. |
| Overlay | Tunnel0, 10.4.0.0/24, GRE multipoint, tunnel key 100, NHRP authentication HALDEN, NHRP Phase 3 |
| Crypto | IKEv2 with pre-shared key Halden-Dmvpn-2026, ESP AES-256 and SHA-256 in transport mode |
| Hubs | HUB1 192.0.2.2 is primary and HUB2 192.0.2.6 is backup. Every spoke registers with both, all the time. |
| Routing | EIGRP named HALDEN, AS 100. Hubs send spokes only 10.0.0.0/8, and spokes are stubs. Hellos 5 seconds, hold 15. |
| Tunnel MTU | 1400 bytes on every overlay interface, and TCP MSS 1360 |
| Spoke to spoke | Direct, after the first packets, through NHRP shortcuts. Encrypted spoke to spoke as well. |
Addressing
| Device | Interface | Address | Connects to |
|---|---|---|---|
| ISP | Lo0 | 10.255.4.254/32 | |
| ISP | Gi0/0 | 192.0.2.1/30 | HUB1 Gi0/0 |
| ISP | Gi0/1 | 192.0.2.5/30 | HUB2 Gi0/0 |
| ISP | Gi0/2 | 192.0.2.9/30 | SP1 Gi0/0 |
| ISP | Gi0/3 | 192.0.2.13/30 | SP2 Gi0/0 |
| ISP | Gi0/4 | 192.0.2.17/30 | SP3 Gi0/0 |
| ISP | Gi0/5 | 192.0.2.21/30 | SP4 Gi0/0 |
| ISP | Gi0/6 | 192.0.2.25/30 | PARTNER Gi0/0 |
| ISP | Gi0/7 | 192.0.2.29/30 | DC1 Gi0/2 |
| HUB1 | Gi0/0 | 192.0.2.2/30 | ISP Gi0/0 |
| HUB1 | Lo0 | 10.255.4.1/32 | |
| HUB1 | Tunnel0 | 10.4.0.1/24 | |
| HUB1 | Gi0/1 | 10.4.101.1/30 | DC1 Gi0/0 |
| HUB2 | Gi0/0 | 192.0.2.6/30 | ISP Gi0/1 |
| HUB2 | Lo0 | 10.255.4.2/32 | |
| HUB2 | Tunnel0 | 10.4.0.2/24 | |
| HUB2 | Gi0/1 | 10.4.102.1/30 | DC1 Gi0/1 |
| DC1 | Lo0 | 10.255.4.100/32 | |
| DC1 | Lo10 | 10.4.100.1/24, 2001:db8:4:100::1/64 | |
| DC1 | Gi0/0 | 10.4.101.2/30 | HUB1 Gi0/1 |
| DC1 | Gi0/1 | 10.4.102.2/30 | HUB2 Gi0/1 |
| DC1 | Gi0/2 | 192.0.2.30/30 | ISP Gi0/7 |
| SP1 | Gi0/0 | 192.0.2.10/30 | ISP Gi0/2 |
| SP1 | Lo0 | 10.255.4.11/32 | |
| SP1 | Lo10 | 10.4.11.1/24, 2001:db8:4:11::1/64 | |
| SP1 | Tunnel0 | 10.4.0.11/24 | |
| SP2 | Gi0/0 | 192.0.2.14/30 | ISP Gi0/3 |
| SP2 | Lo0 | 10.255.4.12/32 | |
| SP2 | Lo10 | 10.4.12.1/24, 2001:db8:4:12::1/64 | |
| SP2 | Tunnel0 | 10.4.0.12/24 | |
| SP3 | Gi0/0 | 192.0.2.18/30 | ISP Gi0/4 |
| SP3 | Lo0 | 10.255.4.13/32 | |
| SP3 | Lo10 | 10.4.13.1/24, 2001:db8:4:13::1/64 | |
| SP3 | Tunnel0 | 10.4.0.13/24 | |
| SP4 | Gi0/0 | 192.0.2.22/30 | ISP Gi0/5 |
| SP4 | Lo0 | 10.255.4.14/32 | |
| SP4 | Lo10 | 10.4.14.1/24, 2001:db8:4:14::1/64 | |
| SP4 | Tunnel0 | 10.4.0.14/24 | |
| PARTNER | Lo10 | 10.4.200.1/24 | |
| PARTNER | Gi0/0 | 192.0.2.26/30 | ISP Gi0/6 |
Rules for the whole lab
- Fix each fault where it is. Don't rebuild tunnels, rename objects or remove the VRF design, and don't reload anything.
- ISP is outside your control.
- Crypto may not be turned off, even for a test. Nothing in the global routing table may route to a public address, apart from the two host routes for the partner tunnel.
- Don't remove the
LAB-FAIL-*applets.
What you’ll practise
- Troubleshoot DMVPN from the underlay up: IKEv2, IPsec, GRE, NHRP and the routing protocol
- Prove spoke-to-spoke shortcuts and dual-hub failover by what traffic does
- Find latent faults that only appear when the primary hub fails
- Build a static GRE tunnel with keepalives, and an IPv6 overlay on the same DMVPN
The tasks
- Section 1: Tickets 1 to 4: bring every site onto the overlay (30 points)
- Section 2: Tickets 5 to 7: make Phase 3 work (24 points)
- Section 3: Tickets 8 and 9: faults nobody has seen yet (16 points)
- Section 4: Static GRE to the partner (12 points)
- Section 5: An IPv6 overlay (18 points)
The Lab Tasks pane has the scenario, these tasks, the checks to run, hints and the full solution.
Topology and devices
The network as it opens in EVE-NG, drawn from the lab’s own canvas, and the image each node runs.
Every link, as text
- ISP
Gi0/0to HUB1Gi0/0 - ISP
Gi0/1to HUB2Gi0/0 - ISP
Gi0/2to SP1Gi0/0 - ISP
Gi0/3to SP2Gi0/0 - ISP
Gi0/4to SP3Gi0/0 - ISP
Gi0/5to SP4Gi0/0 - ISP
Gi0/6to PARTNERGi0/0 - ISP
Gi0/7to DC1Gi0/2 - HUB1
Gi0/1to DC1Gi0/0 - HUB2
Gi0/1to DC1Gi0/1
| Node | Device | Image in Add Node | Set to |
|---|---|---|---|
| ISP | Cisco IOSv | vios-adventerprisek9-m.spa.159-3.m9 | 1 vCPU, 512 MB |
| HUB1 | Cisco IOSv | vios-adventerprisek9-m.spa.159-3.m9 | 1 vCPU, 512 MB |
| HUB2 | Cisco IOSv | vios-adventerprisek9-m.spa.159-3.m9 | 1 vCPU, 512 MB |
| DC1 | Cisco IOSv | vios-adventerprisek9-m.spa.159-3.m9 | 1 vCPU, 512 MB |
| SP1 | Cisco IOSv | vios-adventerprisek9-m.spa.159-3.m9 | 1 vCPU, 512 MB |
| SP2 | Cisco IOSv | vios-adventerprisek9-m.spa.159-3.m9 | 1 vCPU, 512 MB |
| SP3 | Cisco IOSv | vios-adventerprisek9-m.spa.159-3.m9 | 1 vCPU, 512 MB |
| SP4 | Cisco IOSv | vios-adventerprisek9-m.spa.159-3.m9 | 1 vCPU, 512 MB |
| PARTNER | Cisco IOSv | vios-adventerprisek9-m.spa.159-3.m9 | 1 vCPU, 512 MB |
How to open this lab
Sign in to your workspace at
lab.cloudlabbox.com.In File Manager, open Shared, then the folder
CCIE Enterprise Infrastructure practice guided labs.Open
04 - DMVPN Phase 3 dual-hub fault huntand start its nodes, then open Lab Tasks for the scenario and the steps.
The lab is read-only. You can start the nodes and configure the devices, but changes to the topology can’t be saved. To change it, rebuild it in your own folder.
More CCIE Enterprise Infrastructure practice labs
All CCIE Enterprise Infrastructure practice labs
Related labs in other courses
CloudLabBox is independent of Cisco. The lab follows exam topics and isn’t vendor training material; vendor names identify the exams and the images in the workspace.