RKE (Rancher Kubernetes Engine) is a CNCF-certified Kubernetes distribution that runs entirely within Docker containers. The guide will show how create a cluster using RKE.
Installing rke
- Install docker on the instances that will be part of the kubernetes cluster.
- Download the latest available version of RKE binary.
- Copy the RKE binary to a directory that exists in the $PATH and rename it to rke.
- Provide the permission to rke binary as executable.
- Test if it is executable.
Deploying the cluster
For a non minimal cluster, RKE requires cluster.yml to deploy the cluster. Be aware that address and internal_address keys exist in the file. If there's an FIP, it is possible to add its value to address and the private IP address to internal_address. Otherwise, ignore internal_address and set address to the private IP address. Add hostname_override key if you would like to name the Kubernetes nodes else RKE will set their name according to what is present in address.
nodes:
- address: <Private IP address>
user: <user-name of the account in the machine>
role:
- controlplane
- etcd
hostname_override: <kubernetes node name>
- address: <FIP>
user: <user-name of the account in the machine>
role:
- worker
labels:
app: ingress
internal_address: <Private IP address>
hostname_override: <kubernetes node name>
- address: <Private IP address>
user: <user-name of the account in the machine>
role:
- worker
hostname_override: <kubernetes node name>
network:
plugin: canal
options:
canal_flannel_backend_type: vxlan
monitoring:
provider: metrics-server
ingress:
provider: none |
NOTE: There is an option to enable NGINX ingress controller (it is the default option if not specified). However, it allows traffic only within the cluster. Therefore, disable it if you would like to manage ingress from outside the cluster. Later, the instructions show how to deploy your NGINX ingress controller with ingress management from the internet.
- After filling the necessary information in the cluster.yml file, run rke to deploy the cluster.
rke up --config cluster.yml --ssh-agent-auth ~/.ssh/id_rsa |
NOTE: --ssh-agent-auth flag is set when instances require key authentication to be accessed.
RKE will create a kubernetes configuration file called kube_config_cluster.yml so it is necessary to create a copy of this file in the directory where kubectl will use it.
mkdir .kube
cp kube_config_cluster.yml .kube/config |
Deploying In-Cluster Rancher server:
There are two implementations of NGINX ingress controller: One provided by Kubernetes and the other by NGINX. Each one has separate annotations when Ingress resource is deployed. Rancher requires Kubernetes NGINX ingress controller so the following steps to deploy both the ingress controller and Rancher were taken:
kubectl label node k8s-worker-1 kubernetes.io/role=ingress --overwrite |
- create values.yml file for the helm chart of NGINX ingress controller.
controller:
hostNetwork: true
nodeSelector:
kubernetes.io/role: ingress
service:
externalIPs: [131.154.96.63]
type: NodePort
nodePorts:
http: 31811
https: 30506 |
- Install helm chart of NGINX ingress controller.
helm upgrade --install ingress-nginx ingress-nginx \
--repo https://kubernetes.github.io/ingress-nginx \
--namespace ingress-nginx --create-namespace \
--values values.yml |
The output of pods and services should be the following:
ubuntu@k8s-master:~$ kubectl get pods -n ingress-nginx
NAME READY STATUS RESTARTS AGE
ingress-nginx-controller-cb87575f5-7j5zf 1/1 Running 0 3d1h
ubuntu@k8s-master:~$ kubectl get svc -n ingress-nginx
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
ingress-nginx-controller NodePort 10.43.200.65 131.154.96.63 80:31811/TCP,443:30506/TCP 3d1h
ingress-nginx-controller-admission ClusterIP 10.43.60.16 <none> 443/TCP 3d1h |
- Install cert-manager with helm. This tool will manage the certificates for rancher when it's deployed.
- Use helm to deploy rancher. Disable deploying the ingress resource because helm values do not offer changes to the ingress resource before deploying it (customization is required in this case). Therefore use the following command. you can set the website sub domain by changing hostname in the helm command below.
kubectl create namespace cattle-system
helm install rancher rancher-latest/rancher --namespace cattle-system --set hostname=rancher.my.org --set bootstrapPassword=admin --set ingress.enabled=false |
- create a rancher-ingress.yml file to deploy the customized ingress resource of rancher
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
annotations:
cert-manager.io/issuer: rancher
cert-manager.io/issuer-kind: Issuer
meta.helm.sh/release-name: rancher
meta.helm.sh/release-namespace: cattle-system
nginx.ingress.kubernetes.io/proxy-connect-timeout: "30"
nginx.ingress.kubernetes.io/proxy-read-timeout: "1800"
nginx.ingress.kubernetes.io/proxy-send-timeout: "1800"
generation: 2
labels:
app: rancher
app.kubernetes.io/managed-by: Helm
chart: rancher-2.6.3
heritage: Helm
release: rancher
name: rancher
namespace: cattle-system
spec:
ingressClassName: nginx
rules:
- host: rancher.my.org
http:
paths:
- backend:
service:
name: rancher
port:
number: 80
pathType: ImplementationSpecific
tls:
- hosts:
- rancher.my.org
secretName: tls-rancher-ingress |
- After deploying the ingress resource, it should have the external IP address.
ubuntu@k8s-master:~$ kubectl get ingress -A
NAMESPACE NAME CLASS HOSTS ADDRESS PORTS AGE
cattle-system rancher nginx rancher.my.org 131.154.96.63 80, 443 3d1h |
Add the line below to the end of /etc/hosts file on the personal device.
<Floating IP> <hostname>
# example
# 131.154.96.63 rancher.my.org |
Open your browser and visit Rancher.
