Skip to content

Example for L2LS Fabric

Introduction

This example includes and describes all the AVD files used to build a Layer 2 Leaf Spine (L2LS) fabric with the following nodes:

  • Two spine nodes
  • Four leaf nodes

The network fabric in this example is layer 2; an external firewall (FW) or layer 3 (L3) device will handle routing. Later, in this example, we will discuss adding L3 routing to the spines. But first, we will focus on defining the fabric variables to build this L2LS Topology. Before we start, we must ensure we have installed AVD with the requirements covered in the Installation & Requirements section.

The example is meant as a starting foundation. You may build more advanced fabrics based on this design. To keep things simple, the Arista eAPI will be used to communicate with the switches.

Info

The configurations may also be applied with CloudVision with a few updates to your playbook and Ansible variables.

Installation & Requirements

  1. Install AVD - Installation guide found here.
  2. Install Ansible module requirements - Instructions found here.
  3. Run the following playbook to copy the Getting Started examples to your working directory.
# current working directory: ~/ansible-avd-examples
ansible-playbook arista.avd.install_examples

The output will show something similar to the following. If not, please ensure that AVD and all requirements are correctly installed.

 ~/ansible-avd-examples# ansible-playbook arista.avd.install_examples

PLAY [Install Examples]***************************************************************************************************************************************************************************************************************************************************************

TASK [Copy all examples to ~/ansible-avd-examples]*****************************************************************************************************************************************************
changed: [localhost]

PLAY RECAP
****************************************************************************************************************************************************************************************************************************************************************************
localhost                  : ok=1    changed=1    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

After the playbook has run successfully, the following directory structure will be created.

ansible-avd-examples/     (directory where playbook was run)
  ├── l2ls-fabric/
    ├── documentation/
    ├── group_vars/
    ├── images/
    ├── intended/
    ├── switch-basic-configurations/
    ├── ansible.cfg
    ├── build.yml
    ├── deploy.yml
    ├── inventory.yml
    └── README.md (this document)
Info

If the content of any file in the example is modified and the playbook is rerun, the file will not be overwritten. However, if any file in the example is deleted and the playbook is rerun, the file will be re-created.

Design Overview

Physical L2LS Topology

The drawing below shows the physical topology used in this example. The interface assignment shown here are referenced across the entire example, so keep that in mind if this example must be adapted to a different topology.

Figure: 1

Note

In this example, the FW/L3 Device and individual hosts (A-D) are not managed by AVD, but the switch ports connecting to these devices are.

Basic EOS Switch Configuration

Basic connectivity between the Ansible controller host and the switches must be established before Ansible can be used to deploy configurations. The following should be configured on all switches:

  • Switch Hostname
  • IP enabled interface
  • Username and Password defined
  • Management eAPI enabled
Info

When using vEOS/cEOS virtual switches, Management0 or Management1 is used. When using hardware switches, Management1 is used. The included basic switch configurations may need to be adjusted for your environment.

Below is the basic configuration file for SPINE1:

!
no aaa root
!
username admin privilege 15 role network-admin secret sha512 $6$eucN5ngreuExDgwS$xnD7T8jO..GBDX0DUlp.hn.W7yW94xTjSanqgaQGBzPIhDAsyAl9N4oScHvOMvf07uVBFI4mKMxwdVEUVKgY/.
!
hostname SPINE1
!
vrf instance MGMT
!
management api http-commands
   no shutdown
   !
   vrf MGMT
      no shutdown
!
interface Management0
   vrf MGMT
   ip address 172.100.100.101/24
!
ip routing
no ip routing vrf MGMT
!
ip route vrf MGMT 0.0.0.0/0 172.100.100.1
!
management ssh
   vrf MGMT
      no shutdown
!

Ansible Inventory

Now that we understand the physical L2LS topology, we must create the Ansible inventory that represents this topology. The following is a textual and graphical representation of the Ansible inventory group variables and naming scheme used in this example:

- DC1
  - DC1_FABRIC
    - DC1_SPINES
    - DC1_LEAFS
  - DC1_NETWORK_SERVICES
    - DC1_SPINES
    - DC1_LEAFS
  - DC1_NETWORK_PORTS
    - DC1_SPINES
    - DC1_LEAFS

DC1 represents the highest level within the hierarchy. Ansible variables defined at this level will be applied to all nodes in the fabric. Ansible groups have parent-and-child relationships. For example, both DC1_SPINES and DC1_LEAFS are children of DC1_FABRIC. Groups of Groups are possible and allow variables to be shared at any level within the hierarchy. For example, DC1_NETWORK_SERVICES is a group with two other groups defined as children: DC1_SPINES and DC1_LEAFS. The same applies to the group named DC1_NETWORK_PORTS. You will see these groups listed at the bottom of the inventory file.

This naming convention makes it possible to extend anything quickly but can be changed based on your preferences. The names of all groups and hosts must be unique.

Figure: 2

inventory.yml

The below inventory file represents two spines and four leafs. The nodes are defined under the groups DC1_SPINES and DC1_LEAFS, respectively. We apply group variables (group_vars) to these groups to define their functionality and configurations.

The hostnames specified in the inventory must exist either in DNS or in the hosts file on your Ansible host to allow successful name lookup and be able to reach the switches directly. A successful ping from the Ansible host to each inventory host verifies name resolution (e.g., ping SPINE1).

Alternatively, if DNS is unavailable, define the ansible_host variable as an IP address for each device.

# inventory.yml
DC1:
  children:
    DC1_FABRIC:
      children:
        DC1_SPINES:
          hosts:
            SPINE1:
              ansible_host: 172.100.100.101
            SPINE2:
              ansible_host: 172.100.100.102
        DC1_LEAFS:
          hosts:
            LEAF1:
              ansible_host: 172.100.100.105
            LEAF2:
              ansible_host: 172.100.100.106
            LEAF3:
              ansible_host: 172.100.100.107
            LEAF4:
              ansible_host: 172.100.100.108
    DC1_NETWORK_SERVICES:
      children:
        DC1_LEAFS:
        DC1_SPINES:
    DC1_NETWORK_PORTS:
      children:
        DC1_LEAFS:
        DC1_SPINES:

AVD Fabric Variables

To apply AVD variables to the nodes in the fabric, we make use of Ansible group_vars. How and where you define the variables is your choice. The group_vars table below is one example of AVD fabric variables.

group_vars/ Description
DC1.yml Global settings for all devices
DC1_FABRIC.yml Fabric, Topology, and Device settings
DC1_SPINES.yml Device type for spines
DC1_LEAFS.yml Device type for leafs
DC1_NETWORK_SERVICES.yml VLANs
DC1_NETWORK_PORTS.yml Port Profiles and Connected Endpoint settings

The tabs below show the Ansible group_vars used in this example.

At the top level (DC1), the following variables are defined in group_vars/DC1.yml. These Ansible variables apply to all fabric nodes and are a common place to set AAA, users, NTP, and management interface settings. Update local_users and passwords for your environment.

You can create a sha512_password by creating a username and password on a switch and copy/paste it within double quotes here.

---
### group_vars/DC1.yml

aaa_authentication:
  policies:
    local:
      allow_nopassword: true

# local users
local_users:
  # Username with no password configured
  - name: arista
    privilege: 15
    role: network-admin
    no_password: true

  # Username with a password
  - name: admin
    privilege: 15
    role: network-admin
    sha512_password: "$6$eucN5ngreuExDgwS$xnD7T8jO..GBDX0DUlp.hn.W7yW94xTjSanqgaQGBzPIhDAsyAl9N4oScHvOMvf07uVBFI4mKMxwdVEUVKgY/."

# OOB Management network default gateway
mgmt_gateway: 172.100.100.1
mgmt_interface: Management0

# dns servers.
name_servers:
  - 8.8.4.4
  - 8.8.8.8

# NTP Servers IP or DNS name, first NTP server will be preferred, and sourced from Management VRF
ntp:
  servers:
    - name: time.google.com
      preferred: true
      vrf: MGMT
    - name: pool.ntp.org
      vrf: MGMT

# Establish exec/enable role when logging in to switch
aaa_authorization:
  exec:
    default: local

At the fabric level (DC1_FABRIC), the following variables are defined in group_vars/DC1_FABRIC.yml. In addition, the fabric name, design type (l2ls), spine and leaf defaults, ansible authentication, and interface links are defined at this level. Other variables you must supply include spanning-tree mode, priority, and an MLAG IP pool.

Variables applied under the node key type (spine/leaf) defaults section are inherited by nodes under each type. These variables may be overwritten under the node itself.

The spine interface used by a particular leaf is defined from the leaf’s perspective with a variable called uplink_switch_interfaces. For example, LEAF2 has a unique variable uplink_switch_interfaces: [Ethernet2, Ethernet2] defined. This means that LEAF2 is connected to SPINE1’s Ethernet2 and SPINE2’s Ethernet2 interface.

---
### group_vars/DC1_FABRIC.yml

# Set the Fabric Name - must match an Ansible Inventory Group
fabric_name: DC1_FABRIC

# Set Design Type to l2ls
design:
  type: l2ls

# Ansible connectivity definitions
# eAPI connectivity via HTTPS is specified (as opposed to CLI via SSH)
ansible_connection: ansible.netcommon.httpapi
# Specifies that we are indeed using Arista EOS
ansible_network_os: arista.eos.eos
# This user/password must exist on the switches to enable Ansible access
ansible_user: admin
ansible_password: admin
# User escalation (to enter enable mode)
ansible_become: true
ansible_become_method: enable
# Use SSL (HTTPS)
ansible_httpapi_use_ssl: true
# Do not try to validate certs
ansible_httpapi_validate_certs: false

# Spine Switches (L2 only)
spine:
  defaults:
    platform: cEOS-LAB
    spanning_tree_mode: mstp
    spanning_tree_priority: 4096
    mlag_peer_ipv4_pool: 192.168.0.0/24
    mlag_interfaces: [Ethernet47, Ethernet48]
  node_groups:
    - group: SPINES
      nodes:
        - name: SPINE1
          id: 1
          mgmt_ip: 172.100.100.101/24
        - name: SPINE2
          id: 2
          mgmt_ip: 172.100.100.102/24

# Leaf Switches
leaf:
  defaults:
    platform: cEOS-LAB
    mlag_peer_ipv4_pool: 192.168.0.0/24
    uplink_switches: [SPINE1, SPINE2]
    uplink_interfaces: [Ethernet1, Ethernet2]
    mlag_interfaces: [Ethernet47, Ethernet48]
    spanning_tree_mode: mstp
    spanning_tree_priority: 16384

  node_groups:
    - group: RACK1
      mlag: true
      filter:
        tags: [bluezone, greenzone]
      nodes:
        - name: LEAF1
          id: 1
          mgmt_ip: 172.100.100.105/24
          uplink_switch_interfaces: [Ethernet1, Ethernet1]
        - name: LEAF2
          id: 2
          mgmt_ip: 172.100.100.106/24
          uplink_switch_interfaces: [Ethernet2, Ethernet2]
    - group: RACK2
      mlag: true
      filter:
        tags: [bluezone, orangezone]
      nodes:
        - name: LEAF3
          id: 3
          mgmt_ip: 172.100.100.107/24
          uplink_switch_interfaces: [Ethernet3, Ethernet3]
        - name: LEAF4
          id: 4
          mgmt_ip: 172.100.100.108/24
          uplink_switch_interfaces: [Ethernet4, Ethernet4]

#### Override for vEOS/cEOS Lab Caveats ####
p2p_uplinks_mtu: 1500

# Documentation
eos_designs_documentation:
  connected_endpoints: true

In an L2LS design, there are two types of spine nodes: spine and l3spine. In AVD, the node type defines the functionality and the EOS CLI configuration to be generated. For an L2LS design, we will use node type: spine. Later, we will add routing to the spines by changing the node type to l3spine.

---
### group_vars/DC1_SPINES.yml

type: spine     # Must be either spine|l3spine

In an L2LS design, we have one type of leaf node: leaf.

---
### group_vars/DC1_LEAFS.yml

type: leaf     # Must be leaf

You add VLANs to the fabric by updating the group_vars/DC1_NETWORK_SERVICES.yml. Each VLAN will be given a name and a list of tags. The tags filter the VLAN to specific leaf Pairs. These variables are applied to the spine and leaf nodes since they are a part of this group.

---
### group_vars/DC1_NETWORK_SERVICES.yml

tenants:
  - name: MY_FABRIC
    l2vlans:
      - id: 10
        name: 'BLUE-NET'
        tags: [bluezone]
      - id: 20
        name: 'GREEN-NET'
        tags: [greenzone]
      - id: 30
        name: 'ORANGE-NET'
        tags: [orangezone]

Our fabric would only be complete by connecting some devices to it. We define connected endpoints and port profiles in group_vars/DC1_NETWORKS_PORTS.yml. Each endpoint’s adapter defines which switch port(s) and port profile to use. In our example, we have four hosts and a firewall connected to the fabric. The connected endpoints keys are used for logical separation and apply to interface descriptions. These variables are applied to the spine and leaf nodes since they are a part of this inventory group.

---
### group_vars/DC1_NETWORK_PORTS.yml

connected_endpoints_keys:
  - key: servers
    type: server
  - key: firewalls
    type: firewall
  - key: routers
    type: router

port_profiles:
  - profile: PP-DEFAULTS
    spanning_tree_portfast: edge
  - profile: PP-BLUE
    mode: access
    vlans: "10"
    parent_profile: PP-DEFAULTS
  - profile: PP-GREEN
    mode: access
    vlans: "20"
    parent_profile: PP-DEFAULTS
  - profile: PP-ORANGE
    mode: access
    vlans: "30"
    parent_profile: PP-DEFAULTS
  - profile: PP-FIREWALL
    mode: trunk
    vlans: "10,20,30"

servers:
  - name: HostA
    rack: POD1
    adapters:
      - endpoint_ports: [Eth1]
        switch_ports: [Ethernet3]
        switches: [LEAF1]
        profile: PP-BLUE
  - name: HostB
    rack: POD1
    adapters:
      - endpoint_ports: [Eth1]
        switch_ports: [Ethernet3]
        switches: [LEAF2]
        profile: PP-GREEN
  - name: HostC
    rack: POD2
    adapters:
      - endpoint_ports: [Eth1]
        switch_ports: [Ethernet3]
        switches: [LEAF3]
        profile: PP-BLUE
  - name: Host2
    rack: POD2
    adapters:
      - endpoint_ports: [Eth1]
        switch_ports: [Ethernet3]
        switches: [LEAF4]
        profile: PP-ORANGE

firewalls:

  - name: FIREWALL
    adapters:
      - endpoint_ports: [Eth1, Eth2]
        switch_ports: [Ethernet5, Ethernet5]
        switches: [SPINE1, SPINE2]
        profile: PP-FIREWALL
        port_channel:
          mode: active

The Playbooks

Now that we have defined all of our Ansible variables (AVD inputs), it is time to generate some configs. To make things simple, we provide two playbooks. One playbook will allow you to build and view EOS CLI intended configurations per device. The second playbook has an additional task to deploy the configurations to your switches. The playbooks are provided in the tabs below. The playbook is straightforward as it imports two AVD roles: eos_designs and eos_cli_config_gen, which do all the heavy lifting. Combining these two roles produces recommended configurations that follow Arista Design Guides.

---
# build.yml

- name: Build Configs
  hosts: DC1_FABRIC
  gather_facts: false
  tasks:

    - name: Generate AVD Structured Configurations and Fabric Documentation
      ansible.builtin.import_role:
        name: arista.avd.eos_designs

    - name: Generate Device Configurations and Documentation
      ansible.builtin.import_role:
        name: arista.avd.eos_cli_config_gen
---
# deploy.yml

- name: Build and Deploy Configs
  hosts: DC1_FABRIC
  gather_facts: false
  tasks:

    - name: Generate AVD Structured Configurations and Fabric Documentation
      ansible.builtin.import_role:
        name: arista.avd.eos_designs

    - name: Generate Device Configurations and Documentation
      ansible.builtin.import_role:
        name: arista.avd.eos_cli_config_gen

    - name: Deploy Configurations to Devices
      ansible.builtin.import_role:
        name: arista.avd.eos_config_deploy_eapi

Playbook Run

To build the configuration files, run the playbook called build.yml.

### Build configurations
ansible-playbook playbooks/build.yml

After the playbook run finishes, EOS CLI intended configuration files were written to intended/configs.

To build and deploy the configurations to your switches, run the playbook called deploy.yml. This assumes that your Ansible host has access and authentication rights to the switches. Those auth variables were defined in DC1_FABRIC.yml.

### Build configurations & Push Configs to switches
ansible-playbook playbooks/deploy.yml

EOS Intended Configurations

Your configuration files should be similar to these.

!RANCID-CONTENT-TYPE: arista
!
vlan internal order ascending range 1006 1199
!
transceiver qsfp default-mode 4x10G
!
service routing protocols model multi-agent
!
hostname SPINE1
ip name-server vrf MGMT 8.8.4.4
ip name-server vrf MGMT 8.8.8.8
!
ntp server vrf MGMT pool.ntp.org
ntp server vrf MGMT time.google.com prefer
!
spanning-tree mode mstp
no spanning-tree vlan-id 4094
spanning-tree mst 0 priority 4096
!
aaa authentication policy local allow-nopassword-remote-login
aaa authorization exec default local
!
no enable password
no aaa root
!
username admin privilege 15 role network-admin secret sha512 $6$eucN5ngreuExDgwS$xnD7T8jO..GBDX0DUlp.hn.W7yW94xTjSanqgaQGBzPIhDAsyAl9N4oScHvOMvf07uVBFI4mKMxwdVEUVKgY/.
username arista privilege 15 role network-admin nopassword
!
vlan 10
   name BLUE-NET
!
vlan 20
   name GREEN-NET
!
vlan 30
   name ORANGE-NET
!
vlan 4094
   name MLAG_PEER
   trunk group MLAG
!
vrf instance MGMT
!
interface Port-Channel1
   description RACK1_Po1
   no shutdown
   switchport
   switchport trunk allowed vlan 10,20
   switchport mode trunk
   mlag 1
!
interface Port-Channel3
   description RACK2_Po1
   no shutdown
   switchport
   switchport trunk allowed vlan 10,30
   switchport mode trunk
   mlag 3
!
interface Port-Channel5
   description FIREWALL
   no shutdown
   switchport
   switchport trunk allowed vlan 10,20,30
   switchport mode trunk
   mlag 5
!
interface Port-Channel47
   description MLAG_PEER_SPINE2_Po47
   no shutdown
   switchport
   switchport mode trunk
   switchport trunk group MLAG
!
interface Ethernet1
   description LEAF1_Ethernet1
   no shutdown
   channel-group 1 mode active
!
interface Ethernet2
   description LEAF2_Ethernet1
   no shutdown
   channel-group 1 mode active
!
interface Ethernet3
   description LEAF3_Ethernet1
   no shutdown
   channel-group 3 mode active
!
interface Ethernet4
   description LEAF4_Ethernet1
   no shutdown
   channel-group 3 mode active
!
interface Ethernet5
   description FIREWALL_Eth1
   no shutdown
   channel-group 5 mode active
!
interface Ethernet47
   description MLAG_PEER_SPINE2_Ethernet47
   no shutdown
   channel-group 47 mode active
!
interface Ethernet48
   description MLAG_PEER_SPINE2_Ethernet48
   no shutdown
   channel-group 47 mode active
!
interface Management0
   description oob_management
   no shutdown
   vrf MGMT
   ip address 172.100.100.101/24
!
interface Vlan4094
   description MLAG_PEER
   no shutdown
   mtu 1500
   no autostate
   ip address 192.168.0.0/31
no ip routing vrf MGMT
!
mlag configuration
   domain-id SPINES
   local-interface Vlan4094
   peer-address 192.168.0.1
   peer-link Port-Channel47
   reload-delay mlag 300
   reload-delay non-mlag 330
!
ip route vrf MGMT 0.0.0.0/0 172.100.100.1
!
management api http-commands
   protocol https
   no shutdown
   !
   vrf MGMT
      no shutdown
!
end
!RANCID-CONTENT-TYPE: arista
!
vlan internal order ascending range 1006 1199
!
transceiver qsfp default-mode 4x10G
!
service routing protocols model multi-agent
!
hostname SPINE2
ip name-server vrf MGMT 8.8.4.4
ip name-server vrf MGMT 8.8.8.8
!
ntp server vrf MGMT pool.ntp.org
ntp server vrf MGMT time.google.com prefer
!
spanning-tree mode mstp
no spanning-tree vlan-id 4094
spanning-tree mst 0 priority 4096
!
aaa authentication policy local allow-nopassword-remote-login
aaa authorization exec default local
!
no enable password
no aaa root
!
username admin privilege 15 role network-admin secret sha512 $6$eucN5ngreuExDgwS$xnD7T8jO..GBDX0DUlp.hn.W7yW94xTjSanqgaQGBzPIhDAsyAl9N4oScHvOMvf07uVBFI4mKMxwdVEUVKgY/.
username arista privilege 15 role network-admin nopassword
!
vlan 10
   name BLUE-NET
!
vlan 20
   name GREEN-NET
!
vlan 30
   name ORANGE-NET
!
vlan 4094
   name MLAG_PEER
   trunk group MLAG
!
vrf instance MGMT
!
interface Port-Channel1
   description RACK1_Po1
   no shutdown
   switchport
   switchport trunk allowed vlan 10,20
   switchport mode trunk
   mlag 1
!
interface Port-Channel3
   description RACK2_Po1
   no shutdown
   switchport
   switchport trunk allowed vlan 10,30
   switchport mode trunk
   mlag 3
!
interface Port-Channel5
   description FIREWALL
   no shutdown
   switchport
   switchport trunk allowed vlan 10,20,30
   switchport mode trunk
   mlag 5
!
interface Port-Channel47
   description MLAG_PEER_SPINE1_Po47
   no shutdown
   switchport
   switchport mode trunk
   switchport trunk group MLAG
!
interface Ethernet1
   description LEAF1_Ethernet2
   no shutdown
   channel-group 1 mode active
!
interface Ethernet2
   description LEAF2_Ethernet2
   no shutdown
   channel-group 1 mode active
!
interface Ethernet3
   description LEAF3_Ethernet2
   no shutdown
   channel-group 3 mode active
!
interface Ethernet4
   description LEAF4_Ethernet2
   no shutdown
   channel-group 3 mode active
!
interface Ethernet5
   description FIREWALL_Eth2
   no shutdown
   channel-group 5 mode active
!
interface Ethernet47
   description MLAG_PEER_SPINE1_Ethernet47
   no shutdown
   channel-group 47 mode active
!
interface Ethernet48
   description MLAG_PEER_SPINE1_Ethernet48
   no shutdown
   channel-group 47 mode active
!
interface Management0
   description oob_management
   no shutdown
   vrf MGMT
   ip address 172.100.100.102/24
!
interface Vlan4094
   description MLAG_PEER
   no shutdown
   mtu 1500
   no autostate
   ip address 192.168.0.1/31
no ip routing vrf MGMT
!
mlag configuration
   domain-id SPINES
   local-interface Vlan4094
   peer-address 192.168.0.0
   peer-link Port-Channel47
   reload-delay mlag 300
   reload-delay non-mlag 330
!
ip route vrf MGMT 0.0.0.0/0 172.100.100.1
!
management api http-commands
   protocol https
   no shutdown
   !
   vrf MGMT
      no shutdown
!
end
!RANCID-CONTENT-TYPE: arista
!
vlan internal order ascending range 1006 1199
!
transceiver qsfp default-mode 4x10G
!
service routing protocols model multi-agent
!
hostname LEAF1
ip name-server vrf MGMT 8.8.4.4
ip name-server vrf MGMT 8.8.8.8
!
ntp server vrf MGMT pool.ntp.org
ntp server vrf MGMT time.google.com prefer
!
spanning-tree mode mstp
no spanning-tree vlan-id 4094
spanning-tree mst 0 priority 16384
!
aaa authentication policy local allow-nopassword-remote-login
aaa authorization exec default local
!
no enable password
no aaa root
!
username admin privilege 15 role network-admin secret sha512 $6$eucN5ngreuExDgwS$xnD7T8jO..GBDX0DUlp.hn.W7yW94xTjSanqgaQGBzPIhDAsyAl9N4oScHvOMvf07uVBFI4mKMxwdVEUVKgY/.
username arista privilege 15 role network-admin nopassword
!
vlan 10
   name BLUE-NET
!
vlan 20
   name GREEN-NET
!
vlan 4094
   name MLAG_PEER
   trunk group MLAG
!
vrf instance MGMT
!
interface Port-Channel1
   description SPINES_Po1
   no shutdown
   switchport
   switchport trunk allowed vlan 10,20
   switchport mode trunk
   mlag 1
!
interface Port-Channel47
   description MLAG_PEER_LEAF2_Po47
   no shutdown
   switchport
   switchport mode trunk
   switchport trunk group MLAG
!
interface Ethernet1
   description SPINE1_Ethernet1
   no shutdown
   channel-group 1 mode active
!
interface Ethernet2
   description SPINE2_Ethernet1
   no shutdown
   channel-group 1 mode active
!
interface Ethernet3
   description HostA_Eth1
   no shutdown
   switchport access vlan 10
   switchport mode access
   switchport
   spanning-tree portfast
!
interface Ethernet47
   description MLAG_PEER_LEAF2_Ethernet47
   no shutdown
   channel-group 47 mode active
!
interface Ethernet48
   description MLAG_PEER_LEAF2_Ethernet48
   no shutdown
   channel-group 47 mode active
!
interface Management0
   description oob_management
   no shutdown
   vrf MGMT
   ip address 172.100.100.105/24
!
interface Vlan4094
   description MLAG_PEER
   no shutdown
   mtu 1500
   no autostate
   ip address 192.168.0.0/31
no ip routing vrf MGMT
!
mlag configuration
   domain-id RACK1
   local-interface Vlan4094
   peer-address 192.168.0.1
   peer-link Port-Channel47
   reload-delay mlag 300
   reload-delay non-mlag 330
!
ip route vrf MGMT 0.0.0.0/0 172.100.100.1
!
management api http-commands
   protocol https
   no shutdown
   !
   vrf MGMT
      no shutdown
!
end
!RANCID-CONTENT-TYPE: arista
!
vlan internal order ascending range 1006 1199
!
transceiver qsfp default-mode 4x10G
!
service routing protocols model multi-agent
!
hostname LEAF2
ip name-server vrf MGMT 8.8.4.4
ip name-server vrf MGMT 8.8.8.8
!
ntp server vrf MGMT pool.ntp.org
ntp server vrf MGMT time.google.com prefer
!
spanning-tree mode mstp
no spanning-tree vlan-id 4094
spanning-tree mst 0 priority 16384
!
aaa authentication policy local allow-nopassword-remote-login
aaa authorization exec default local
!
no enable password
no aaa root
!
username admin privilege 15 role network-admin secret sha512 $6$eucN5ngreuExDgwS$xnD7T8jO..GBDX0DUlp.hn.W7yW94xTjSanqgaQGBzPIhDAsyAl9N4oScHvOMvf07uVBFI4mKMxwdVEUVKgY/.
username arista privilege 15 role network-admin nopassword
!
vlan 10
   name BLUE-NET
!
vlan 20
   name GREEN-NET
!
vlan 4094
   name MLAG_PEER
   trunk group MLAG
!
vrf instance MGMT
!
interface Port-Channel1
   description SPINES_Po1
   no shutdown
   switchport
   switchport trunk allowed vlan 10,20
   switchport mode trunk
   mlag 1
!
interface Port-Channel47
   description MLAG_PEER_LEAF1_Po47
   no shutdown
   switchport
   switchport mode trunk
   switchport trunk group MLAG
!
interface Ethernet1
   description SPINE1_Ethernet2
   no shutdown
   channel-group 1 mode active
!
interface Ethernet2
   description SPINE2_Ethernet2
   no shutdown
   channel-group 1 mode active
!
interface Ethernet3
   description HostB_Eth1
   no shutdown
   switchport access vlan 20
   switchport mode access
   switchport
   spanning-tree portfast
!
interface Ethernet47
   description MLAG_PEER_LEAF1_Ethernet47
   no shutdown
   channel-group 47 mode active
!
interface Ethernet48
   description MLAG_PEER_LEAF1_Ethernet48
   no shutdown
   channel-group 47 mode active
!
interface Management0
   description oob_management
   no shutdown
   vrf MGMT
   ip address 172.100.100.106/24
!
interface Vlan4094
   description MLAG_PEER
   no shutdown
   mtu 1500
   no autostate
   ip address 192.168.0.1/31
no ip routing vrf MGMT
!
mlag configuration
   domain-id RACK1
   local-interface Vlan4094
   peer-address 192.168.0.0
   peer-link Port-Channel47
   reload-delay mlag 300
   reload-delay non-mlag 330
!
ip route vrf MGMT 0.0.0.0/0 172.100.100.1
!
management api http-commands
   protocol https
   no shutdown
   !
   vrf MGMT
      no shutdown
!
end
!RANCID-CONTENT-TYPE: arista
!
vlan internal order ascending range 1006 1199
!
transceiver qsfp default-mode 4x10G
!
service routing protocols model multi-agent
!
hostname LEAF3
ip name-server vrf MGMT 8.8.4.4
ip name-server vrf MGMT 8.8.8.8
!
ntp server vrf MGMT pool.ntp.org
ntp server vrf MGMT time.google.com prefer
!
spanning-tree mode mstp
no spanning-tree vlan-id 4094
spanning-tree mst 0 priority 16384
!
aaa authentication policy local allow-nopassword-remote-login
aaa authorization exec default local
!
no enable password
no aaa root
!
username admin privilege 15 role network-admin secret sha512 $6$eucN5ngreuExDgwS$xnD7T8jO..GBDX0DUlp.hn.W7yW94xTjSanqgaQGBzPIhDAsyAl9N4oScHvOMvf07uVBFI4mKMxwdVEUVKgY/.
username arista privilege 15 role network-admin nopassword
!
vlan 10
   name BLUE-NET
!
vlan 30
   name ORANGE-NET
!
vlan 4094
   name MLAG_PEER
   trunk group MLAG
!
vrf instance MGMT
!
interface Port-Channel1
   description SPINES_Po3
   no shutdown
   switchport
   switchport trunk allowed vlan 10,30
   switchport mode trunk
   mlag 1
!
interface Port-Channel47
   description MLAG_PEER_LEAF4_Po47
   no shutdown
   switchport
   switchport mode trunk
   switchport trunk group MLAG
!
interface Ethernet1
   description SPINE1_Ethernet3
   no shutdown
   channel-group 1 mode active
!
interface Ethernet2
   description SPINE2_Ethernet3
   no shutdown
   channel-group 1 mode active
!
interface Ethernet3
   description HostC_Eth1
   no shutdown
   switchport access vlan 10
   switchport mode access
   switchport
   spanning-tree portfast
!
interface Ethernet47
   description MLAG_PEER_LEAF4_Ethernet47
   no shutdown
   channel-group 47 mode active
!
interface Ethernet48
   description MLAG_PEER_LEAF4_Ethernet48
   no shutdown
   channel-group 47 mode active
!
interface Management0
   description oob_management
   no shutdown
   vrf MGMT
   ip address 172.100.100.107/24
!
interface Vlan4094
   description MLAG_PEER
   no shutdown
   mtu 1500
   no autostate
   ip address 192.168.0.4/31
no ip routing vrf MGMT
!
mlag configuration
   domain-id RACK2
   local-interface Vlan4094
   peer-address 192.168.0.5
   peer-link Port-Channel47
   reload-delay mlag 300
   reload-delay non-mlag 330
!
ip route vrf MGMT 0.0.0.0/0 172.100.100.1
!
management api http-commands
   protocol https
   no shutdown
   !
   vrf MGMT
      no shutdown
!
end
!RANCID-CONTENT-TYPE: arista
!
vlan internal order ascending range 1006 1199
!
transceiver qsfp default-mode 4x10G
!
service routing protocols model multi-agent
!
hostname LEAF4
ip name-server vrf MGMT 8.8.4.4
ip name-server vrf MGMT 8.8.8.8
!
ntp server vrf MGMT pool.ntp.org
ntp server vrf MGMT time.google.com prefer
!
spanning-tree mode mstp
no spanning-tree vlan-id 4094
spanning-tree mst 0 priority 16384
!
aaa authentication policy local allow-nopassword-remote-login
aaa authorization exec default local
!
no enable password
no aaa root
!
username admin privilege 15 role network-admin secret sha512 $6$eucN5ngreuExDgwS$xnD7T8jO..GBDX0DUlp.hn.W7yW94xTjSanqgaQGBzPIhDAsyAl9N4oScHvOMvf07uVBFI4mKMxwdVEUVKgY/.
username arista privilege 15 role network-admin nopassword
!
vlan 10
   name BLUE-NET
!
vlan 30
   name ORANGE-NET
!
vlan 4094
   name MLAG_PEER
   trunk group MLAG
!
vrf instance MGMT
!
interface Port-Channel1
   description SPINES_Po3
   no shutdown
   switchport
   switchport trunk allowed vlan 10,30
   switchport mode trunk
   mlag 1
!
interface Port-Channel47
   description MLAG_PEER_LEAF3_Po47
   no shutdown
   switchport
   switchport mode trunk
   switchport trunk group MLAG
!
interface Ethernet1
   description SPINE1_Ethernet4
   no shutdown
   channel-group 1 mode active
!
interface Ethernet2
   description SPINE2_Ethernet4
   no shutdown
   channel-group 1 mode active
!
interface Ethernet3
   description Host2_Eth1
   no shutdown
   switchport access vlan 30
   switchport mode access
   switchport
   spanning-tree portfast
!
interface Ethernet47
   description MLAG_PEER_LEAF3_Ethernet47
   no shutdown
   channel-group 47 mode active
!
interface Ethernet48
   description MLAG_PEER_LEAF3_Ethernet48
   no shutdown
   channel-group 47 mode active
!
interface Management0
   description oob_management
   no shutdown
   vrf MGMT
   ip address 172.100.100.108/24
!
interface Vlan4094
   description MLAG_PEER
   no shutdown
   mtu 1500
   no autostate
   ip address 192.168.0.5/31
no ip routing vrf MGMT
!
mlag configuration
   domain-id RACK2
   local-interface Vlan4094
   peer-address 192.168.0.4
   peer-link Port-Channel47
   reload-delay mlag 300
   reload-delay non-mlag 330
!
ip route vrf MGMT 0.0.0.0/0 172.100.100.1
!
management api http-commands
   protocol https
   no shutdown
   !
   vrf MGMT
      no shutdown
!
end

Add Routing to Spines

Our example used an external L3/FW Device to route between subnets. This is very typical in a Layer 2 only environment. To route on the spines, we remove the L3/FW device from the topology and create the SVIs on the spines. The updated topology is shown below.

Note

The spine type has been changed to l3spine.

Figure: 3

The following group_vars need updating to enable L3 routing on the spines.

  • DC1_SPINES.yml
  • DC1_FABRIC.yml
  • DC1_NETWORK_SERVICES.yml

The updated changes are noted in the tabs below.

Update type to l3spine. This makes it a routing device.

---
### group_vars/DC1_SPINES.yml

type: l3spine

Update with the following changes and additions.

  • Change the node key spine to l3spine to match the node type set previously in DC1_SPINES.yml
  • Add loopback_ipv4_pool
  • Add mlag_peer_l3_ipv4_pool
  • Add virtual_router_mac_address

Update DC1_FABRIC.yml with the following recommended settings. Use your own IP pools.

# Node Key must be l3spine to match type
l3spine:
  defaults:
    platform: cEOS-LAB
    spanning_tree_mode: mstp
    spanning_tree_priority: 4096
    # Loopback is used to generate a router-id
    loopback_ipv4_pool: 1.1.1.0/24
    mlag_peer_ipv4_pool: 192.168.0.0/24
    # Needed for L3 peering across the MLAG Trunk
    mlag_peer_l3_ipv4_pool: 10.1.1.0/24
    # Used for SVI Virtual MAC address
    virtual_router_mac_address: 00:1c:73:00:dc:01
    mlag_interfaces: [Ethernet47, Ethernet48]

Update Network Services to use L3 SVIs.

Note

To create L3 SVIs on the spines, we need to utilize an L3 VRF. In our case, we will use the default VRF. MY_FABRIC is simply a tenant name for organizing VRFs and SVIs.

tenants:
  - name: MY_FABRIC
    vrfs:
      - name: default
        svis:
          - id: 10
            name: 'BLUE-NET'
            tags: [bluezone]
            enabled: true
            ip_virtual_router_addresses:
              - 10.10.10.1
            nodes:
              - node: SPINE1
                ip_address: 10.10.10.2/24
              - node: SPINE2
                ip_address: 10.10.10.3/24
          - id: 20
            name: 'GREEN-NET'
            tags: [greenzone]
            enabled: true
            ip_virtual_router_addresses:
              - 10.20.20.1
            nodes:
              - node: SPINE1
                ip_address: 10.20.20.2/24
              - node: SPINE2
                ip_address: 10.20.20.3/24
          - id: 30
            name: 'ORANGE-NET'
            tags: [orangezone]
            enabled: true
            ip_virtual_router_addresses:
              - 10.30.30.1
            nodes:
              - node: SPINE1
                ip_address: 10.30.30.2/24
              - node: SPINE2
                ip_address: 10.30.30.3/24

Now rerun your playbook and build the new configurations. The intended/configs for the spines will have been updated with L3 SVIs.

ansible-playbook playbooks/build.yml

If you wish to deploy these changes, then run the deploy playbook.

ansible-playbook playbooks/deploy.yml

Next steps

Try building your topology.


Last update: June 1, 2023