Skip to content

Latest commit

 

History

History

rke

README.md

Configure

Copy then edit gateway, domain...

cp group_vars/all.yml.template group_vars/all.yml

RancherOS

RancherOS is the OS which will run Kubernetes Cluster.
On each worker|controller nodes

  • Boot RancherOS iso
  • Configure network
    To configure static ip
sudo ros config set rancher.network.interfaces.eth0.address X.X.X.X/24
sudo ros config set rancher.network.interfaces.eth0.gateway X.X.X.Y
sudo ros config set rancher.network.interfaces.eth0.dhcp false
sudo system-docker restart network

Ansible
To be able to run ansible on nodes, it needs python.

  • Change to ubuntu console with
sudo ros console switch ubuntu -f
sudo apt update
sudo apt install python -y
  • Change default rancher password
sudo passwd rancher

Then on your host

# Test ansible connectivity
ansible workers:controllers -m ping -u rancher -k
ansible-playbook 00-configure.yml -k
# enter rancher password
# it will fail with unreachable status du to reboot, that's normal

Load balancer

TODO

RKE

RKE automates the setup process of Kubernetes cluster.
Generate RKE config based on ansible hosts file.

ansible-playbook 01-local.yml

Then run rke manually with

cd rke
./rke up

Test with

# Install kubeconfig
mkdir -p ~/.kube
# edit kube_config_cluster.yml to set load-balancer hostname, then merge kubeconfig
KUBECONFIG=~/.kube/config:./kube_config_cluster.yml kubectl config view 
KUBECONFIG=~/.kube/config:./kube_config_cluster.yml kubectl config view --raw > ~/.kube/config
kubectl get nodes

Persistent Volumes

ee rook/rook#3923 (comment)
Workers nodes needs an extra unformated disk to configure CephFS.

# deploy rook
kubectl create -f files/rook/common.yaml
kubectl create -f files/rook/operator.yaml
# verify that rook-ceph-operator is running
watch kubectl -n rook-ceph get pod
# then
kubectl create -f files/rook/cluster.yaml

# deploy toolbox
kubectl create -f files/rook/toolbox.yml
# verify that toolbox is running
kubectl -n rook-ceph get pod -l "app=rook-ceph-tools"
# run the toolbox
kubectl -n rook-ceph exec -it (kubectl -n rook-ceph get pod -l "app=rook-ceph-tools" -o jsonpath='{.items[0].metadata.name}') -- bash
# check ceph
ceph status # health should be "ok"
ceph osd status
ceph df
rados df

Rancher

Rancher is the tool to manage Kubernetes Clusters on a WebUI.
Helm is the Kubernetes package manager.

Install helm package manager on your host and add rancher repository

helm repo add rancher-latest https://releases.rancher.com/server-charts/latest
# Create a namespace
kubectl create namespace cattle-system

Install cert-manager, then

helm install rancher rancher-latest/rancher \
  --namespace cattle-system \
  --set hostname=rancher.my.org

Create hosts file entry or DNS to match rancher hostname.
Then access to https://rancher.my.org