cluster-network-addons-operator

README

Cluster Network Addons Operator

This operator can be used to deploy additional networking components on top of Kubernetes cluster.

Configuration

Configuration of desired network addons is done using NetworkAddonsConfig object:

apiVersion: networkaddonsoperator.network.kubevirt.io/v1
kind: NetworkAddonsConfig
metadata:
  name: cluster
spec:
  multus: {}
  linuxBridge: {}
  kubeMacPool: {}
  nmstate: {}
  ovs: {}
  macvtap: {}
  imagePullPolicy: Always

Multus

The operator allows administrator to deploy multi-network Multus plugin. It is done using multus attribute.

apiVersion: networkaddonsoperator.network.kubevirt.io/v1
kind: NetworkAddonsConfig
metadata:
  name: cluster
spec:
  multus: {}

Additionally, container image used to deliver this plugin can be set using MULTUS_IMAGE environment variable in operator deployment manifest.

Linux Bridge

The operator allows administrator to deploy Linux Bridge CNI plugin simply by adding linuxBridge attribute to NetworkAddonsConfig.

apiVersion: networkaddonsoperator.network.kubevirt.io/v1
kind: NetworkAddonsConfig
metadata:
  name: cluster
spec:
  linuxBridge: {}

Additionally, container image used to deliver this plugin can be set using LINUX_BRIDGE_IMAGE environment variable in operator deployment manifest.

The bridge marker image used to deliver a bridge marker detecting the availability of linux bridges on nodes can be set using the LINUX_BRIDGE_MARKER_IMAGE environment variable in operator deployment manifest.

Configure bridge on node

Following snippets can be used to configure linux bridge on your node.

# create the bridge using NetworkManager
nmcli con add type bridge ifname br10

# allow traffic to go through the bridge between pods
iptables -I FORWARD 1 -i br10 -j ACCEPT

Kubemacpool

The operator allows administrator to deploy the Kubemacpool. This project allow to allocate mac addresses from a pool to secondary interfaces using Network Plumbing Working Group de-facto standard.

Note: Administrator can specify a requested range, if the range is not requested a random range will be provided. This random range spans from 02:XX:XX:00:00:00 to 02:XX:XX:FF:FF:FF, where 02 makes the address local unicast and XX:XX is a random prefix.

apiVersion: networkaddonsoperator.network.kubevirt.io/v1
kind: NetworkAddonsConfig
metadata:
  name: cluster
spec:
  kubeMacPool:
   rangeStart: "02:00:00:00:00:00"
   rangeEnd: "FD:FF:FF:FF:FF:FF"

NMState

Note: This feature is experimental. NMState is unstable and its API may change.

The operator allows the administrator to deploy the NMState State Controller as a daemonset across all of one's nodes. This project manages host networking settings in a declarative manner. The networking state is described by a pre-defined schema. Reporting of current state and changes to it (desired state) both conform to the schema. NMState is aimed to satisfy enterprise needs to manage host networking through a northbound declarative API and multi provider support on the southbound. NetworkManager acts as the main (and currently the only) provider supported.

This component can be enabled by adding nmstate section to the NetworkAddonsConfig.

apiVersion: networkaddonsoperator.network.kubevirt.io/v1
kind: NetworkAddonsConfig
metadata:
  name: cluster
spec:
  nmstate: {}

It communicate with a NetworkManager instance running on the node using D-Bus. Make sure that NetworkManager is installed and running on each node.

yum install NetworkManager
systemctl start NetworkManager

Open vSwitch

The operator allows administrator to deploy OVS CNI plugin simply by adding ovs attribute to NetworkAddonsConfig. Please note that in order to use this plugin, openvswitch have to be up and running at nodes.

apiVersion: networkaddonsoperator.network.kubevirt.io/v1
kind: NetworkAddonsConfig
metadata:
  name: cluster
spec:
  ovs: {}

Macvtap

Note: This feature is experimental. Macvtap-cni is unstable and its API may change.

The operator allows the administrator to deploy the macvtap CNI plugin, simply by adding macvtap attribute to NetworkAddonsConfig.

apiVersion: networkaddonsoperator.network.kubevirt.io/v1
kind: NetworkAddonsConfig
metadata:
  name: cluster
spec:
  macvtap: {}

Macvtap-cni must be explicitly configured by the administrator, indicating the interfaces on top of which logical networks can be created.

A simple example on how to do so, the user must deploy a ConfigMap, such as in this example.

Currently, this configuration is not dynamic.

Image Pull Policy

Administrator can specify image pull policy for deployed components. Default is IfNotPresent.

apiVersion: networkaddonsoperator.network.kubevirt.io/v1
kind: NetworkAddonsConfig
metadata:
  name: cluster
spec:
  imagePullPolicy: Always

Self Signed Certificates Configuration

Administrator can specify webhook self signed certificates configuration for deployed components. Default is caRotateInterval: 168h, caOverlapInterval: 24h, certRotateInterval: 24h, certOverlapInterval: 8h

apiVersion: networkaddonsoperator.network.kubevirt.io/v1
kind: NetworkAddonsConfig
metadata:
  name: cluster
spec:
  selfSignConfiguration:
    caRotateInterval: 168h
    caOverlapInterval: 24h
    certRotateInterval: 24h
    certOverlapInterval: 8h

The selfSignConfiguration parameters has to be all or none set, setting some of them fails at validation. They have to conform to golang time.Duration string format. Additionally the following checks are done at validation:

• caRotateInterval >= caOverlapInterval && caRotateInterval >= certRotateInterval && certRotateInterval >= certOverlapInterval

This parameters are consumed by Kubemacpool and Kubernetes-nmstate components.

Placement Configuration

CNAO deploys two component categories: infra and workload. Workload components manage node configuration and need to be scheduled on the nodes where actual user workload is scheduled. Infra components provide cluster-wide service and do not need to be running on the same nodes as user workload.

Administrator can specify placement preferences for deployed infra and workload components by defining affinity, nodeSelector and tolerations.

By default, infra components are scheduled on control-plane nodes and workload components are scheduled on all nodes. To adjust this behaviour, provide custom placementConfiguration to the NetworkAddonsConfig.

In the following example, nodeAffinity is used to schedule infra components to control-plane nodes and nodeSelector to schedule workloads on worker nodes. Note that worker nodes need to be labeled with node-role.kubernetes.io/worker label.

apiVersion: networkaddonsoperator.network.kubevirt.io/v1
kind: NetworkAddonsConfig
metadata:
  name: cluster
spec:
  placementConfiguration:
    infra:
      affinity:
        nodeAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
            nodeSelectorTerms:
            - matchExpressions:
              - key: node-role.kubernetes.io/control-plane
                operator: Exists
    workloads:
      nodeSelector:
        node-role.kubernetes.io/worker: ""

Deployment

First install the operator itself:

kubectl apply -f https://github.com/kubevirt/cluster-network-addons-operator/releases/download/v0.65.2/namespace.yaml
kubectl apply -f https://github.com/kubevirt/cluster-network-addons-operator/releases/download/v0.65.2/network-addons-config.crd.yaml
kubectl apply -f https://github.com/kubevirt/cluster-network-addons-operator/releases/download/v0.65.2/operator.yaml

Then you need to create a configuration for the operator example CR:

kubectl apply -f https://github.com/kubevirt/cluster-network-addons-operator/releases/download/v0.65.2/network-addons-config-example.cr.yaml

Finally you can wait for the operator to finish deployment:

kubectl wait networkaddonsconfig cluster --for condition=Available

In case something failed, you can find the error in the NetworkAddonsConfig Status field:

kubectl get networkaddonsconfig cluster -o yaml

You can follow the deployment state through events produced in the default namespace:

kubectl get events

Events will be produced whenever the deployment is applied, configured or failed. The expected events are:

Event type	Reason
Progressing	When operator had started deploying the components
Failed	When one or more components failed to deploy
Available	When all components finished to deploy
Modified	When the configuration was modified or applied for the first time

For more information about the configuration format check configuring section.

Upgrades

Starting with version 0.42.0, this operator supports upgrades to any newer version. If you wish to upgrade, remove old operator (operator.yaml) and install new, operands will remain available during the operator's downtime.

Development

Make sure you have Docker >= 17.05 installed.

# run code validation and unit tests
make check

# perform auto-formatting on the source code (if not done by your IDE)
make fmt

# generate source code for API
make gen-k8s

# build images (uses multi-stage builds and therefore requires Docker >= 17.05)
make docker-build

# or build only a specific image
make docker-build-operator
make docker-build-registry

# bring up a local cluster with Kubernetes
make cluster-up

# bridge up a local cluster with kubernetes 1.21
export KUBEVIRT_PROVIDER=k8s-1.21
make cluster-up

# build images and push them to the local cluster
make cluster-operator-push

# install operator on the local cluster
make cluster-operator-install

# run workflow e2e tests on the cluster, requires cluster with installed operator,
# workflow covers deployment of operands
make test/e2e/workflow

# run lifecycle e2e tests on the cluster, requires cluster without operator installed,
# lifecycle covers deployment of operator itself and its upgrades
make test/e2e/lifecycle

# access kubernetes API on the cluster
./cluster/kubectl.sh get nodes

# ssh into the cluster's node
./cluster/cli.sh ssh node01

# clean up all resources created by the operator from the cluster
make cluster-clean

# delete the cluster
make cluster-down

For developing at an external cluster:

export KUBEVIRT_PROVIDER=external

export KUBECONFIG=[path to external's cluster kubeconfig]

# This is the registry used to push and pull the dev image
# it has to be accessible by the external cluster
export DEV_IMAGE_REGISTRY=quay.io/$USER

# Then is possible to follow normal dev flow

make cluster-operator-push
make cluster-operator-install
make test/e2e/workflow
make test/e2e/lifecycle

Releasing

1. Checkout a public branch
2. Call make prepare-patch|minor|major and prepare release notes
3. Open a new PR
4. Once the PR is merged, create a new release in GitHub and attach new manifests