This guide is intended as a way to get you off the ground, using Multus CNI to create Kubernetes pods with multiple interfaces. If you're already using Multus and need more detail, see the comprehensive usage guide. This document is a quickstart and a getting started guide in one, intended for your first run-through of Multus CNI.
We'll first install Multus CNI, and then we'll setup some configurations so that you can see how multiple interfaces are created for pods.
Two things we'll refer to a number of times through this document are:
- "Default network" -- This is your pod-to-pod network. This is how pods communicate among one another in your cluster, how they have connectivity. Generally speaking, this is presented as the interface named
eth0. This interface is always attached to your pods, so that they can have connectivity among themselves. We'll add interfaces in addition to this. - "CRDs" -- Custom Resource Definitions. Custom Resources are a way that the Kubernetes API is extended. We use these here to store some information that Multus can read. Primarily, we use these to store the configurations for each of the additional interfaces that are attached to your pods.
Our installation method requires that you first have installed Kubernetes and have configured a default network -- that is, a CNI plugin that's used for your pod-to-pod connectivity.
We support Kubernetes versions that Kubernetes community supports. Please see Supported versions in Kubernetes document.
To install Kubernetes, you may decide to use kubeadm, or potentially kubespray.
After installing Kubernetes, you must install a default network CNI plugin. If you're using kubeadm, refer to the "Installing a pod network add-on" section in the kubeadm documentation. If it's your first time, we generally recommend using Flannel for the sake of simplicity.
Alternatively, for advanced use cases, for installing Multus and a default network plugin at the same time, you may refer to the Kubernetes Network Plumbing Group's Reference Deployments.
To verify that you default network is ready, you may list your Kubernetes nodes with:
kubectl get nodes
In the case that your default network is ready you will see the STATUS column also switch to Ready for each node.
NAME STATUS ROLES AGE VERSION
master-0 Ready master 1h v1.17.1
master-1 Ready master 1h v1.17.1
master-2 Ready master 1h v1.17.1
Our recommended quickstart method to deploy Multus is to deploy using a Daemonset (a method of running pods on each nodes in your cluster), this spins up pods which install a Multus binary and configure Multus for usage.
We'll apply a YAML file with kubectl from this repo, which installs the Multus components.
Recommended installation:
kubectl apply -f https://raw.githubusercontent.com/k8snetworkplumbingwg/multus-cni/master/deployments/multus-daemonset-thick.yml
See the thick plugin docs for more information about this architecture.
Alternatively, you may install the thin-plugin with:
kubectl apply -f https://raw.githubusercontent.com/k8snetworkplumbingwg/multus-cni/master/deployments/multus-daemonset.yml
- Starts a Multus daemonset, this runs a pod on each node which places a Multus binary on each node in
/opt/cni/bin - Reads the lexicographically (alphabetically) first configuration file in
/etc/cni/net.d, and creates a new configuration file for Multus on each node as/etc/cni/net.d/00-multus.conf, this configuration is auto-generated and is based on the default network configuration (which is assumed to be the alphabetically first configuration) - Creates a
/etc/cni/net.d/multus.ddirectory on each node with authentication information for Multus to access the Kubernetes API.
Generally, the first step in validating your installation is to ensure that the Multus pods have run without error, you may see an overview of those by looking at:
kubectl get pods --all-namespaces | grep -i multus
You may further validate that it has ran by looking at the /etc/cni/net.d/ directory and ensure that the auto-generated /etc/cni/net.d/00-multus.conf exists corresponding to the alphabetically first configuration file.
The first thing we'll do is create configurations for each of the additional interfaces that we attach to pods. We'll do this by creating Custom Resources. Part of the quickstart installation creates a "CRD" -- a custom resource definition that is the home where we keep these custom resources -- we'll store our configurations for each interface in these.
Each configuration we'll add is a CNI configuration. If you're not familiar with them, let's break them down quickly. Here's an example CNI configuration:
{
"cniVersion": "0.3.0",
"type": "loopback",
"additional": "information"
}
CNI configurations are JSON, and we have a structure here that has a few things we're interested in:
cniVersion: Tells each CNI plugin which version is being used and can give the plugin information if it's using a too late (or too early) version.type: This tells CNI which binary to call on disk. Each CNI plugin is a binary that's called. Typically, these binaries are stored in/opt/cni/binon each node, and CNI executes this binary. In this case we've specified theloopbackbinary (which create a loopback-type network interface). If this is your first time installing Multus, you might want to verify that the plugins that are in the "type" field are actually on disk in the/opt/cni/bindirectory.additional: This field is put here as an example, each CNI plugin can specify whatever configuration parameters they'd like in JSON. These are specific to the binary you're calling in thetypefield.
For an even further example -- take a look at the bridge CNI plugin README which shows additional details.
If you'd like more information about CNI configuration, you can read the entire CNI specification. It might also be useful to look at the CNI reference plugins and see how they're configured.
You do not need to reload or refresh the Kubelets when CNI configurations change. These are read on each creation & deletion of pods. So if you change a configuration, it'll apply the next time a pod is created. Existing pods may need to be restarted if they need the new configuration.
So, we want to create an additional interface. Let's create a macvlan interface for pods to use. We'll create a custom resource that defines the CNI configuration for interfaces.
Note in the following command that there's a kind: NetworkAttachmentDefinition. This is our fancy name for our configuration -- it's a custom extension of Kubernetes that defines how we attach networks to our pods.
Secondarily, note the config field. You'll see that this is a CNI configuration just like we explained earlier.
Lastly but very importantly, note under metadata the name field -- here's where we give this configuration a name, and it's how we tell pods to use this configuration. The name here is macvlan-conf -- as we're creating a configuration for macvlan.
Here's the command to create this example configuration:
cat <<EOF | kubectl create -f -
apiVersion: "k8s.cni.cncf.io/v1"
kind: NetworkAttachmentDefinition
metadata:
name: macvlan-conf
spec:
config: '{
"cniVersion": "0.3.0",
"type": "macvlan",
"master": "eth0",
"mode": "bridge",
"ipam": {
"type": "host-local",
"subnet": "192.168.1.0/24",
"rangeStart": "192.168.1.200",
"rangeEnd": "192.168.1.216",
"routes": [
{ "dst": "0.0.0.0/0" }
],
"gateway": "192.168.1.1"
}
}'
EOF
NOTE: This example uses eth0 as the master parameter, this master parameter should match the interface name on the hosts in your cluster.
You can see which configurations you've created using kubectl here's how you can do that:
kubectl get network-attachment-definitions
You can get more detail by describing them:
kubectl describe network-attachment-definitions macvlan-conf
We're going to create a pod. This will look familiar as any pod you might have created before, but, we'll have a special annotations field -- in this case we'll have an annotation called k8s.v1.cni.cncf.io/networks. This field takes a comma delimited list of the names of your NetworkAttachmentDefinitions as we created above. Note in the command below that we have the annotation of k8s.v1.cni.cncf.io/networks: macvlan-conf where macvlan-conf is the name we used above when we created our configuration.
Let's go ahead and create a pod (that just sleeps for a really long time) with this command:
cat <<EOF | kubectl create -f -
apiVersion: v1
kind: Pod
metadata:
name: samplepod
annotations:
k8s.v1.cni.cncf.io/networks: macvlan-conf
spec:
containers:
- name: samplepod
command: ["/bin/ash", "-c", "trap : TERM INT; sleep infinity & wait"]
image: alpine
EOF
You may now inspect the pod and see what interfaces are attached, like so:
kubectl exec -it samplepod -- ip a
You should note that there are 3 interfaces:
loa loopback interfaceeth0our default networknet1the new interface we created with the macvlan configuration.
For additional confirmation, use kubectl describe pod samplepod and there will be an annotations section, similar to the following:
Annotations: k8s.v1.cni.cncf.io/networks: macvlan-conf
k8s.v1.cni.cncf.io/network-status:
[{
"name": "cbr0",
"ips": [
"10.244.1.73"
],
"default": true,
"dns": {}
},{
"name": "macvlan-conf",
"interface": "net1",
"ips": [
"192.168.1.205"
],
"mac": "86:1d:96:ff:55:0d",
"dns": {}
}]
This metadata tells us that we have two CNI plugins running successfully.
You can add more interfaces to a pod by creating more custom resources and then referring to them in pod's annotation. You can also reuse configurations, so for example, to attach two macvlan interfaces to a pod, you could create a pod like so:
cat <<EOF | kubectl create -f -
apiVersion: v1
kind: Pod
metadata:
name: samplepod
annotations:
k8s.v1.cni.cncf.io/networks: macvlan-conf,macvlan-conf
spec:
containers:
- name: samplepod
command: ["/bin/ash", "-c", "trap : TERM INT; sleep infinity & wait"]
image: alpine
EOF
Note that the annotation now reads k8s.v1.cni.cncf.io/networks: macvlan-conf,macvlan-conf. Where we have the same configuration used twice, separated by a comma.
If you were to create another custom resource with the name foo you could use that such as: k8s.v1.cni.cncf.io/networks: foo,macvlan-conf, and use any number of attachments.