3. Multihoming - Switch Configuration

In this section, you will configure existing VTEPs to participate in multihoming. The commands are identical for Leaf1 (VTEP1) and Leaf2 (VTEP2). For this demonstration, Leaf1 (VTEP1) has been pre-configured. You will only be adding the configuration for Leaf2 (VTEP2). A constant ping between hosts will be used throughout this section to display when connectivity is present. Additional validation is covered in the next section.

Use ping to detect changes in network

Use the next few steps to start a ping between these two tenants: - VTEP1: **Host1-40** (192.168.40.101) - VTEP26: **Host2-40** (192.168.40.102) [[Note: If these two hosts show RED cable connections, they have not yet started. Right click each host icon and select Start from the pull-down menu. Once started, the host connections will be GREEN.]] Leaf2 switch will not initially participate in the ping, but will join the ping once it has been configured for multihoming. Right click **Host1-40.**

Click "Console"

Click on the Host1-40 console that opens. Then press [[Return]] Use credentials **gns3/gns3** to login.

At the **gns3@gns3:\~$** prompt, type **ping 192.168.40.102**[[Return]]

Notice the successful pings between the tenants, without using Leaf2 (not yet configured). Leave the ping session going and viewable while making changes in the next steps.

**Alert:** If pings are returning **Destination Host Unreachable**, continue to the next step. Most likely either **Leaf1** or **Leaf26** is not at a **System is ready** state. One or more may need to be reloaded , which is covered in a later step.

Right-click **Leaf1.** Click "Console"

The ***first time*** a switch console opens after it boots up (but before logging in with credentials), the message at the bottom should show **System is ready.** If the message is not displayed, follow the recommendations in the ***Alert*** below.

**ALERT**: If **System is ready** is not displayed the first time you open the switch console, try the following one at a time until you can log in. 1. Press **Return** and try to login anyway. When near completing bootup, login will still function normally. 2. If you see an error trying to login, wait up to 3 minutes for the **System is ready** message to appear, then try to login again. 3. If after 3 minutes, if you haven't seen the **System is ready** message, right click the switch and click **Reload.** The reload process will take 1 to 5 minutes, at which time you should see the **System is ready** message.

Click on the **Leaf1** console that opens. Then press Enter.

Enter login credentials, **admin/admin123**. At the **admin@Leaf1:\~$** prompt, enter **sonic-cli**[[Return]]

In this step you will shut down the **Leaf1 (Eth 1/1)** port to stop ping traffic going through **Leaf1.** This will allow you to see when traffic through **Leaf2** is obtained. At the **Leaf1#** prompt, enter the following: **configure terminal** **interface Eth 1/1** **shutdown**

The ping from Host1-40 changes to "**Destination Host Unreachable**" indicating the two hosts are no longer able to communicate. Note: The ping will stop immediately, but it can take up to 30 seconds for the message to display.

Configure VTEPs for multihoming

In the next set of steps, you will configure **Leaf2** with EVPN multihoming. Right-click **Leaf2.** Click "Console"

Click inside the **Leaf2** console. Press Return to show prompt.

Enter the following commands to configure multihoming globally on **Leaf2**, and add a startup-delay. The startup delay is the number of seconds multihoming interfaces are kept down during the VTEP bootup process to avoid traffic loss. **config terminal** **evpn esi-multihoming** **startup-delay 300** **exit**

Enter the following commands to configure the required Ethernet segment ID (ES-ID) on each port channel that connects to multihomed VTEPs. ES-ID Type-0, Type-1, and Type-3 are supported (per RFC 7432). A 10-byte Type-0 is used in this example. The existing VLAN 40 is allowed on the port channel and the interface is then activated. **interface PortChannel 1** **evpn ethernet-segment 00:00:00:11:22:33:00:00:00:01** **exit** **system-mac 00:00:00:11:22:33** **switchport trunk allowed vlan add 40** **no shutdown**

Enter the following commands to apply the channel-group (PortChannel 1) configuration to the physical interface (Eth 1/1) that is connected to the tenant Host1-40. Perform a **no shutdown** to bring up the interface. **interface Eth 1/1** **channel-group 1** **no shutdown**

Notice ping responses have resumed between Host1-40 and Host2-40.

This time the traffic is going across **Leaf2** (and the new cable) instead of **Leaf1**. Recall that **Leaf1** interface Eth 1/1 was shut down in step 10.

In the interest of time, Leaf1 was preconfigured with EVPN Multihoming using the same commands that were used for Leaf2 in this lab. In an actual deployment, all participating leafs would need to be configured.

Return to **Leaf1**. Make sure you are still at the **Leaf1(config-if-Eth1/1)#** prompt, and enter "**no shutdown**" to re-enabled the port going from Leaf1 to Host1-40. **Leaf1(config-if-Eth1/1)#** no shutdown

EVPN network traffic for Host1-40 is now going across both Leaf1 and Leaf2 in a multihomed configuration. This demonstrates the added redundancy and increased bandwidth capabilities provided by EVPN multihoming.

Check that traffic is traversing both paths

Be sure the ping session in your Host1-40 console is visible while entering commands in the Leaf1 and Leaf2 consoles.

On the **Leaf2** command line, make sure you are at the **Leaf2(config-if-Eth1/1)#** command prompt for configuring the interface. Enter **shutdown**. **Leaf2(config-if-Eth1/1)#** shutdown

Notice the ping between **Host1-40** and **Host2-40** successfully continues! Traffic for **Host1-40** is now going across **Leaf1** only. This time it is using a multihoming configuration, but in a degraded state (loss of a path), demonstrating the redundancy and failover capabilities of EVPN multihoming. In situations where multiple ports or switches in a fabric become temporarily unresponsive, EVPN multihoming helps ensure continuous connectivity for the end nodes.

On Leaf1, make sure you are at the command prompt for configuring the interface: **Leaf1(config-if-Eth1/1)#.** Enter **shutdown.** **Leaf1(config-if-Eth1/1)#** shutdown Once again, the ping from the **Host1-40** console stops and shows the message **Destination Host Unreachable.**

Note: The ping will stop immediately, but it can take up to 30 seconds for the message to display.

All multihoming paths for Host1-40 are now down, and traffic to and from the host is lost as shown. If more redundancy is desired for increased bandwidth or stronger failover capabilities, additional links may be added between the hosts and VTEPs. Up to four multihoming links may be used for each host.

Verify traffic is load balanced across both paths using show commands

Use the **no shutdown** command on both **Leaf1 and Leaf2** command lines to bring **UP** both multihoming paths: - **Leaf1(config-if-Eth1/1)#** no shutdown - **Leaf2(config-if-Eth1/1)#** no shutdown Check to make sure the ping is showing successful responses before proceeding to the next step.

On the **Leaf1** command line enter **do show interface counters Eth 1/1 | grep Without** Using the **UP arrow** key on the keyboard to automatically retype the command, run it multiple times to see that the number of packets is increasing on the **Leaf1** multihoming interface.

Troubleshooting tip: If "Total Packets Received Without Errors" is not increasing for both Leaf1 and Leaf2, go back to the previous step and verify the **no shutdown** command was ran on both ports.

On the **Leaf2** command line enter the same command **do show interface counters Eth 1/1 | grep Without** Use the **UP arrow** key again to retype the command on Leaf2. Run this command multiple times to see the number of packets increasing on the **Leaf2** multihoming interface.

The **show counters** commands demonstrate Host1-40 traffic is traversing both Leaf1 and Leaf2 simultaneously in a load-balanced multihomed configuration.

Click Close at the top of this guide and proceed to Part 4 - Validate the configuration.