Getting started with MetalLB, the bare metal load balancer for Kubernetes

Need to route traffic on bare-metal Kubernetes clusters? MetalLB is a solution you should consider. As a load balancer that is purpose-built for bare-metal Kubernetes, MetalLB lets you deploy a Kubernetes Service of type LoadBalancer. Then, MetalLB automatically routes traffic for you, using a pool of IP addresses that you define.

Although MetalLB is not the only traffic management solution for bare-metal Kubernetes (you could use an alternative bare-metal load balancer like Porter, or a technique like NodePort), MetalLB offers some standout features. Keep reading for details on why MetalLB is a good choice, how it works, as well as how to get started using it.

Kubernetes MetalLB over NodePort or Ingress

MetalLB isn’t the only way to route traffic in bare-metal Kubernetes clusters. Other methods include:

NodePort, which lets you route traffic from external endpoints to specific node and port configurations inside your cluster. NodePort is easy to set up, but it is subject to many limitations, not least of which is that you can only run one Service on each NodePort. You also have to update your configuration manually if your nodes’ IP assignments change.
An Ingress Service, which exposes a Service to an external URL or other endpoint. This approach is relatively sophisticated, and it gives you a lot of control over how traffic is routed, but it requires you to set up your own ingress controller on all of your nodes.

In the past, these were the main ways to manage traffic for bare-metal clusters. But they were not ideal for the reasons described above.

In comparison, MetalLB offers a load balancing solution for bare metal that is:

Easy to set up: You just have to create a load balancer Service and assign an IP address pool to MetalLB.
Easy to scale: MetalLB automates most aspects of traffic management, and unlike NodePort and ExternalIP, it requires minimal changes to your configurations. It can thus scale to support mappings for dozens or even hundreds of applications, if you need.
Consistent with cloud-based load balancing: Because MetalLB uses LoadBalancer Service configurations rather than specialized routing methods like NodePort, you can configure MetalLB in a very similar fashion to configuring load balancing in any public cloud-based Kubernetes environment. If you’re already familiar with Kubernetes load balancing in AWS, Azure or GCP, then, you’ll probably find MetalLB easy to work with.

For these reasons, MetalLB is the obvious solution for most bare-metal routing and load balancing needs today. Older methods, like NodePort and ExternalIP, are still fully supported in Kubernetes, at least for now. But MetalLB is the simpler and more sophisticated option for most use cases.

Installing MetalLB

The simplicity of MetalLB starts with the installation process. To install MetalLB in most clusters, first enable strict ARP mode by running:

 
	  kubectl edit configmap -n kube-system kube-proxy

And make sure the strictARP value is set to true:

 
	  apiVersion: kubeproxy.config.k8s.io/v1alpha1
kind: KubeProxyConfiguration
mode: "ipvs"
ipvs:
  strictARP: true

Then, install MetalLB with kubectl:

 
	    kubectl apply -f https://raw.githubusercontent.com/metallb/metallb/v0.12.1/manifests/namespace.yaml
  kubectl apply -f https://raw.githubusercontent.com/metallb/metallb/v0.12.1/manifests/metallb.yaml

You can alternatively install MetalLB with Helm, if you prefer:

 
	      helm repo add metallb https://metallb.github.io/metallb
  helm install metallb metallb/metallb

Configuring MetalLB

The easiest way to use MetalLB is to configure networking at layer 2. Under this approach, you simply assign a range of IP addresses to MetalLB. It then automatically assigns them nodes and manages traffic between them and your endpoints.

To define the address pool, open the MetalLB ConfigMap with:

 
	        kubectl edit configmap config -n metallb-system

Then define the address-pools and addresses values as desired. For example, to use the range 192.168.255.1–192.168.255.255, set the configuration to:

 
    apiVersion: v1
    kind: ConfigMap
    metadata:
      namespace: metallb-system
      name: config
    data:
      config:
        address-pools:
    	  - name: default
      	protocol: layer2
      	  addresses:
          - 192.168.255.1-192.168.255.255

After making changes, apply them with:

 
      kubectl rollout restart deployment controller -n metallb-system

Once MetalLB is provisioned with IP addresses, simply define a LoadBalancer Service to route traffic to an application, just as you would in a public cloud environment. For example:

 
          apiVersion: v1
    kind: Service
    metadata:
      name: my-service
    spec:
      selector:
    	app: MyApp
      ports:
    	- protocol: TCP
      	port: 80
      	targetPort: 9376
      clusterIP: 10.0.171.239
      type: LoadBalancer
    status:
      loadBalancer:
    	ingress:
    	- ip: 192.0.2.127

Advanced MetalLB usage

The steps above reflect the simplest way to get started with MetalLB. But there are many more things you can do with the load balancer, especially if you opt for BGP mode instead of using layer 2 configurations. Under BGP mode, you set up peering sessions between each node and your routers, providing more fine-grained control over traffic routing than you would have at layer 2.

The MetalLB documentation offers full details on configuring MetalLB for BGP mode, along with other advanced setup options. Note, too, that MetalLB is one of the load balancers natively available through Palette. Read more about using MetalLB and Palette together in our documentation, or sign up for a free account to see MetalLB and Palette in action.

Tags:

Bare Metal

How to