Skip to content

Quickstart#

Installation#

Getting containerlab is as easy as it gets. Thanks to the quick setup script you can get up and running in a matter of seconds on any RHEL or Debian based OS1.

curl -sL https://containerlab.dev/setup | sudo -E bash -s "all"

To complete installation and enable sudo-less docker command execution, please run newgrp docker or logout and log back in.

Topology definition file#

Once installed, containerlab manages the labs defined in the so-called topology definition, clab files. A user can write a topology definition file from scratch, or look at various lab examples we provided within the containerlab package or explore dozens of labs our community has shared.

In this quickstart we will be using one of the provided labs which consists of Nokia SR Linux and Arista cEOS nodes connected one to another.

The lab topology is defined in the srlceos01.clab.yml file. To make use of this lab example, we need to fetch the clab file:

mkdir ~/clab-quickstart #(1)!
cd ~/clab-quickstart

curl -LO \
https://raw.githubusercontent.com/srl-labs/containerlab/main/lab-examples/srlceos01/srlceos01.clab.yml #(2)!
  1. Create a directory to store the lab definition file.
  2. Download the lab definition file.

Let's have a look at how this lab's topology is defined:

# topology documentation: http://containerlab.dev/lab-examples/srl-ceos/
name: srlceos01

topology:
  nodes:
    srl:
      kind: nokia_srlinux
      image: ghcr.io/nokia/srlinux:24.3.3
    ceos:
      kind: arista_ceos
      image: ceos:4.32.0F

  links:
    - endpoints: ["srl:e1-1", "ceos:eth1"]

A topology definition deep-dive document provides a complete reference of the topology definition syntax. In this quickstart we keep it short, glancing over the key components of the file:

  • Each lab has a name.
  • The lab topology is defined under the topology element.
  • Topology is a set of nodes and links between them.
  • The nodes are always of a certain kind. The kind defines the node configuration and behavior.
  • Containerlab supports a fixed number of kinds. In the example above, the srl and ceos are one of the supported kinds.
  • The actual nodes of the topology are defined in the nodes section which holds a map of node names. In the example above, nodes with names srl and ceos are defined.
  • Node elements must have a kind parameter to indicate which kind this node belongs to. Under the nodes section you typically provide node-specific parameters. This lab uses a node-specific parameters - image.
  • nodes are interconnected with links. Each link is defined by a set of endpoints.

Container image#

One of node's most important properties is the container image that the nodes are defined with. The image name follows the same rules as the images you use with, for example, Docker CLI or Docker Compose.

Image name formats and fully qualified names

There are several forms of how one can write an image name, but each name essentially maps to a fully qualified image name that consists of:

  1. registry - the registry where the image is stored.
  2. organisation - the organisation name in that registry
  3. repository - the repository name
  4. tag - the image tag

Below you will find different image names you can come across when working with containerlab and the corresponding FQDN names they map to:

  1. ghcr.io/nokia/srlinux
    • registry: ghcr.io
    • organisation: nokia
    • repository: srlinux
    • tag: when no explicit tag is set, the implicit latest is used
  2. ceos:4.32.0F
    • registry: when registry is not set, implicit docker.io registry is assumed
    • organisation: when none is set, implicit library organisation is assumed
    • repository: ceos
    • tag: 4.32.0F
  3. prom/prometheus:v2.47
    • registry: when registry is not set, implicit docker.io registry is assumed
    • organisation: prom
    • repository: prometheus
    • tag: v2.47

Our topology file defines two nodes, where each node is defined with its own container image:

  • ghcr.io/nokia/srlinux:24.3.3 - for Nokia SR Linux node
  • ceos:4.32.0F - for Arista cEOS node

When containerlab starts to deploy the lab, it will first check if the images are available locally. Local images can be listed with docker images command.

Containerlab will compare the image names from the topology file with the local images and if the images are not available locally, it will try to pull them from the remote registry.

Images availability

Quickstart lab includes Nokia SR Linux and Arista cEOS images. While Nokia SR Linux is a publicly available image and can be pulled by anyone, its counterpart, Arista cEOS image, is not available in a public registry.

Arista requires its users to register with arista.com before downloading any images. Once you created an account and logged in, go to the software downloads section and download ceos64 tar archive for a given release.

Once downloaded, import the archive with docker:

# import container image and save it under ceos:4.32.0F name
docker import cEOS64-lab-4.32.0F.tar.xz ceos:4.32.0F

Deploying a lab#

Now when we know what a basic topology file consists of, refreshed our knowledge on what container image name is and imported cEOS image, we can proceed with deploying this lab. To keep things easy and guessable, the command to deploy a lab is called deploy.

Doesn't hurt to verify that we have cEOS image imported before we hit the deploy command:

docker images | grep ceos
REPOSITORY                        TAG        IMAGE ID       CREATED         SIZE
ceos                              4.32.0F    40d39e1a92c2   24 hours ago    2.4GB

Remote topology files

Containerlab allows to deploy labs from files located in remote Git repositories and/or HTTP URLs. Check out deploy command documentation for more details.

While you can pre-pull the Nokia SR Linux image, containerlab will do it for you if it's not available locally, handy! So we are ready to deploy:

sudo containerlab deploy # (1)!
  1. deploy command will automatically lookup a file matching the *.clab.y*ml patter to select it.
    If you have several files and want to pick a specific one, use --topo <path> flag.

In no time you will see the summary table with the deployed lab nodes.
The table will show the node name (which equals to container name), node kind, image name and a bunch of other useful information. You can always list the nodes of the lab with containerlab inspect command.

+---+---------------------+--------------+-----------------------+---------------+---------+-----------------+----------------------+
| # |        Name         | Container ID |         Image         |     Kind      |  State  |  IPv4 Address   |     IPv6 Address     |
+---+---------------------+--------------+-----------------------+---------------+---------+-----------------+----------------------+
| 1 | clab-srlceos01-ceos | 6ec1b1367a77 | ceos:4.32.0F          | arista_ceos   | running | 172.20.20.11/24 | 3fff:172:20:20::b/64 |
| 2 | clab-srlceos01-srl  | 6af1e33f4573 | ghcr.io/nokia/srlinux | nokia_srlinux | running | 172.20.20.10/24 | 3fff:172:20:20::a/64 |
+---+---------------------+--------------+-----------------------+---------------+---------+-----------------+----------------------+

Connecting to the nodes#

We know you want to get your hands dirty with the nodes, so let's connect to them. The common way netengs use to interact with network devices is via CLI. With Network OSes you can use SSH to connect to the CLI by either using the management address assigned to the container or a node name:

❯ ssh admin@clab-srlceos01-srl

Using configuration file(s): []
Welcome to the srlinux CLI.
Type 'help' (and press <ENTER>) if you need any help using this.
--{ running }--[  ]--
A:srl#

Note

For each supported kind we document the management interfaces and the ways to leverage them.
For example, srl kind documentation provides the commands to leverage SSH and gNMI interfaces.
ceos kind has its own instructions.

The following tab view aggregates the ways to get CLI access per the lab node:

Since the topology nodes are regular containers, you can connect to them just like to any other container.

# access CLI
docker exec -it clab-srlceos01-srl sr_cli
# access bash
docker exec -it clab-srlceos01-srl bash
# access CLI
docker exec -it clab-srlceos01-ceos Cli
# access bash
docker exec -it clab-srlceos01-ceos bash

Feel free to explore the nodes, configure them, and run your favorite network protocols. If you break something, you can always destroy the lab and start over. Speaking of which...

Destroying a lab#

To remove the lab, use the destroy command that takes a topology file as an argument:

sudo containerlab destroy

What next?#

To get a broader view on the containerlab features and components, refer to the User manual section.

Do not forget to check out the Lab examples section where we provide complete and ready-to-run topology definition files. This is a great starting point to explore containerlab by doing.


  1. For other installation options such as package managers, manual binary downloads or instructions to get containerlab for non-RHEL/Debian distros, refer to the installation guide