Setup KubeEdge
June 29, 2025 · View on GitHub
KubeEdge is to extend native containerized application orchestration capabilities to hosts at Edge. That said, KubeEdge requires Kubernetes cluster running already in the cluster infrastructure. Tunneling between cloudcore and edgecore provides control-plane connectivity beyond internet, so that edge devices behind NAT or different network can join the cluster running in the cloud infrastructure. Kubernetes worker nodes are transparent, so that the same operation can be applied with Kubernetes.
Reference
Kubernetes Compatibility
KubeEdge depends on the Kubernetes version, see more details for KubeEdge Kubernetes Compatibility.
Container Network Interface (CNI)
Although KubeEdge Roadmap mentions that it supports CNI, it still requires the CNI dependent special operation to instantiate the CNI implementation with KubeEdge.
Please see more details for,
KubeEdge community also has been developing edgemesh for next-generation data-plane component, including the support as CNI.
Setup Kubernetes API Server
As explained above, KubeEdge requires Kubernetes cluster running. In other words, Kubernetes API server must be running before KubeEdge installation.
see Setup Kubernetes API Server and Access API-server.
In Kubernetes workloads, it requires one of CNI implementation running to set up the cluster. see Deploy CNI Plugin to start the CNI for Kubernetes. (this CNI can only be used by Kubernetes worker nodes but KubeEdge.)
This is only required to bring the Kubernetes API-server running, because we are going to deploy cloudcore to the same physical node with Kubernetes API-server.
Instead of having CNI deployed to bring the Kubernetes API-server up and running, we are not able to deploy cloudcore to the node since we cannot deploy the containers to any NotReady nodes.
Setup KubeEdge
see more details for https://kubeedge.io/docs/setup/install-with-keadm/
Install keadm
keadm needs to be installed in both cloud and edge nodes.
> wget https://github.com/kubeedge/kubeedge/releases/download/v1.19.1/keadm-v1.19.1-linux-amd64.tar.gz
> tar -zxvf keadm-v1.19.1-linux-amd64.tar.gz
> cp keadm-v1.19.1-linux-amd64/keadm/keadm /usr/local/bin/keadm
> keadm version
version: version.Info{Major:"1", Minor:"19", GitVersion:"v1.19.1", GitCommit:"e676d31ee10ee8a9a17e74b38134a11bdc1d8350", GitTreeState:"clean", BuildDate:"2024-12-02T12:11:18Z", GoVersion:"go1.21.11", Compiler:"gc", Platform:"linux/amd64"}
Cloud Core
- with the configuration, we will deploy the KubeEdge
cloudcoreto Kubernetes master node. Basically master node has the taints not to schedule the pods to keep the system resource for Kubernetes. So we need to remove that taint so that we can deploy thecloudcorepods to the master node. (if this operation is not done,keadm initwill fail withError: timed out waiting for the condition)
> kubectl get nodes -o json | jq '.items[].spec.taints'
[
{
"effect": "NoSchedule",
"key": "node-role.kubernetes.io/control-plane"
}
]
> kubectl taint nodes tomoyafujita node-role.kubernetes.io/control-plane:NoSchedule-
node/tomoyafujita untainted
> kubectl taint nodes --all node-role.kubernetes.io/master-
> kubectl taint nodes --all node-role.kubernetes.io/control-plane-
> kubectl get nodes -o json | jq '.items[].spec.taints'
null
- start KubeEdge
cloudcore.
> keadm init --advertise-address=<YOUR_IP_ADDRESS> --kube-config=/root/.kube/config --kubeedge-version=v1.19.1
Kubernetes version verification passed, KubeEdge v1.19.1 installation will start...
CLOUDCORE started
=========CHART DETAILS=======
Name: cloudcore
LAST DEPLOYED: Mon Feb 3 15:36:04 2025
NAMESPACE: kubeedge
STATUS: deployed
REVISION: 1
> kubectl get all -n kubeedge
NAME READY STATUS RESTARTS AGE
pod/cloudcore-7c54886ddb-kd2vh 1/1 Running 0 42s
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
service/cloudcore ClusterIP 10.111.74.141 <none> 10000/TCP,10001/UDP,10002/TCP,10003/TCP,10004/TCP 42s
NAME DESIRED CURRENT READY UP-TO-DATE AVAILABLE NODE SELECTOR AGE
daemonset.apps/edge-eclipse-mosquitto 0 0 0 0 0 <none> 42s
NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/cloudcore 1/1 1 1 42s
NAME DESIRED CURRENT READY AGE
replicaset.apps/cloudcore-7c54886ddb 1 1 1 42s
Edge Core
- get security token from cloudcore.
> keadm gettoken
30dad7c33966782b1fd0aa21b068d348e2acbf7464339882f10f6ce23d10e0f2.eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJleHAiOjE3Mzg3MTIxNzN9.3NuaxcMlxXaCxV4Bebl54oxisL_VUwCtViX6xjAmIr0
- start KubeEdge
edgecore
> keadm join --cloudcore-ipport=<YOUR_IP_ADDRESS>>:10000 --token=30dad7c33966782b1fd0aa21b068d348e2acbf7464339882f10f6ce23d10e0f2.eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJleHAiOjE3Mzg3MTIxNzN9.3NuaxcMlxXaCxV4Bebl54oxisL_VUwCtViX6xjAmIr0 --kubeedge-version=v1.19.1 --remote-runtime-endpoint=unix:///run/containerd/containerd.sock --cgroupdriver systemd
...<snip>
I0203 15:39:14.636693 990048 join_others.go:273] KubeEdge edgecore is running, For logs visit: journalctl -u edgecore.service -xe
I0203 15:39:24.645991 990048 join.go:94] 9. Install Complete!
> systemctl status edgecore
● edgecore.service
Loaded: loaded (/etc/systemd/system/edgecore.service; enabled; vendor preset: enabled)
Active: active (running) since Mon 2025-02-03 15:39:14 PST; 1min 21s ago
Main PID: 990292 (edgecore)
Tasks: 24 (limit: 18670)
Memory: 33.3M
CPU: 6.113s
CGroup: /system.slice/edgecore.service
└─990292 /usr/local/bin/edgecore
...<snip>
- check cluster nodes.
kubectl get nodes -o wide
KubeEdge Test Deployment
> kubectl apply -f ros2-sample-hostnic.yaml
deployment.apps/ros2-talker-1 created
deployment.apps/ros2-listener-1 created
Break Down KubeEdge
> \rm -rf /etc/kubeedge/*
> kubeadm deprecated reset --force
[reset] Reading configuration from the cluster...