Import Kubeadm Cluster to Rancher Manager

In this article I will show you how to import existing Kubeadm cluster to Rancher Manager. Rancher Manager providing a single web interface to monitor, manage, and operate all cluster. It gives you centralized dashboards, unified RBAC, application catalogs, and consistent monitoring across your entire Kubernetes infrastructure. All without changing how your cluster operates underneath. This is Prerequisites cluster for import to rancher manager.

So, let's get started…

Environment

• RKE2 Cluster and Rancher Manager

Hostname	at-rke2-1 (master node)
Operating System	Ubuntu 22.04 (Jammy)
vCPU	8 (too large for testing)
Memory	12 GB (too large for testing)
Disk	60 GB
Network	172.20.20.65

Hostname	at-rke2-2 (worker node)
Operating System	Ubuntu 22.04 (Jammy)
vCPU	4
Memory	8 GB
Disk	40 GB
Network	172.20.20.66

Hostname	at-rke2-3 (ingress node)
Operating System	Ubuntu 22.04 (Jammy)
vCPU	2
Memory	4 GB
Disk	30 GB
Network	172.20.20.67

• Kubeadm Cluster

Hostname	at-kubeadm (master node)
Operating System	Ubuntu 22.04 (Jammy)
vCPU	4
Memory	8 GB
Disk	60 GB
Network	172.20.20.75

Hostname	at-kubeadm-2 (worker node)
Operating System	Ubuntu 22.04 (Jammy)
vCPU	2
Memory	4 GB
Disk	40 GB
Network	172.20.20.76

Import Kubeadm Cluster to Rancher Manager

1. Mapping hosts

# exec on rke2 cluster nodes
nano /etc/hosts
---
172.20.20.65 at-rke2-1 at-rke2-1.at.lab rancher.at.lab
172.20.20.66 at-rke2-2 at-rke2-2.at.lab
172.20.20.67 at-rke2-3 at-rke2-3.at.lab

172.20.20.75 at-kubeadm kubeadm.at.lab
172.20.20.76 at-kubeadm-2

# exec on kubeadm cluster nodes
nano /etc/hosts
---
172.20.20.65 at-rke2-1 at-rke2-1.at.lab
172.20.20.66 at-rke2-2 at-rke2-2.at.lab
172.20.20.67 at-rke2-3 at-rke2-3.at.lab rancher.at.lab

172.20.20.75 at-kubeadm kubeadm.at.lab
172.20.20.76 at-kubeadm-2

1. Import existing cluster

1. Set name and create import generic cluster

1. Copy registration command

1. Run registration command in kubeadm cluster

# exec on kubeadm master node (at-kubeadm)
curl --insecure -sfL https://rancher.at.lab/v3/import/dsxdbngbwfvj75k6nnlh2zf6jqtp7wl66d9gvv5w7qrhtq7tv4jc5s_c-fjlfv.yaml | kubectl apply -f -

# when error no resolve domain
kubectl edit deployment cattle-cluster-agent -n cattle-system
---
      dnsPolicy: ClusterFirstWithHostNet
      hostNetwork: true

by default replica for the cattle-cluster-agent deployment is 2. so if you just have 1 control-plane/master node, cattle-cluster-agent will trying run 2 pod in that node because affinity, that’s why 1 pod CLBO because conflict port. but you can delete CLBO pod to running in other node (worker)

1. Verification

1. Install and configure kubectx for centralize cluster management

# exec on rke2 master node (at-rke2-1)
git clone https://github.com/ahmetb/kubectx /opt/kubectx
ln -s /opt/kubectx/kubectx /usr/local/bin/kubectx

access https://rancher.at.lab/dashboard/home -> Manage -> click option/triple dot -> Copy KubeConfig to Clipboard

# exec on rke2 master node (at-rke2-1)
nano /root/.kube/config
---
apiVersion: v1
kind: Config
clusters:
- name: "local"
  cluster:
    server: "https://rancher.at.lab/k8s/clusters/local"
    certificate-authority-data: "[REDACTED]"
- name: "kubeadm-cluster"
  cluster:
    server: "https://rancher.at.lab/k8s/clusters/c-fjlfv"
    certificate-authority-data: "[REDACTED]"

users:
- name: "local"
  user:
    token: "[REDACTED]"
- name: "kubeadm-cluster"
  user:
    token: "[REDACTED]"

contexts:
- name: "rke2-cluster"
  context:
    user: "local"
    cluster: "local"
- name: "kubeadm-cluster"
  context:
    user: "kubeadm-cluster"
    cluster: "kubeadm-cluster"

current-context: "rke2-cluster"

# exec on rke2 master node (at-rke2-1)
nano ~/.bashrc
---
#export KUBECONFIG=/etc/rancher/rke2/rke2.yaml
export KUBECONFIG=/root/.kube/config

source ~/.bashrc

# mapping domain to ingress ip because kubectx connect cluster via external/ingress
nano /etc/hosts
---
172.20.20.65 #rancher.at.lab
172.20.20.67 rancher.at.lab

1. Switch cluster with kubectx

# exec on rke2 master node (at-rke2-1)
kubectx
kubectx kubeadm-cluster
kubectl get nodes -o wide

1. Operational test

# exec on rke2 master node (at-rke2-1)
kubectl create deployment nginx-import --image=nginx --replicas=1
kubectl expose deployment nginx-import --port=80 --target-port=80 
kubectl create ingress nginx-import-ingress --class=nginx --rule="nginx-import.at.lab/*=nginx-import:80"

nano /etc/hosts
---
172.20.20.75 nginx-import.at.lab

Thank You.