Kubernetes, an open-source orchestration system, automates the deployment and management of containerized applications. For beginners, the journey into Kubernetes can often start with the daunting question: "Where do I begin?"
In the early days, setting up and managing a Kubernetes cluster was a challenging and time-consuming task. However, with the evolution of Kubernetes, user-friendly solutions have emerged to simplify this process. Among these solutions, Kubespray shines as an invaluable tool.
Kubespray, an open-source solution, facilitates the automated deployment of Kubernetes clusters across nodes. Engineered to be highly customizable, efficient, and lightweight, Kubespray caters to a wide range of requirements, making Kubernetes cluster deployment accessible to all.
Overview of Kubespray
Kubespray is a composition of Ansible playbooks, inventory, provisioning tools, and generic Kubernetes cluster configuration management tasks. In this writeup, I'll demonstrate how to deploy a Kubernetes cluster on 3 nodes (1master and 2 worker nodes) using Kubespray.
While a basic understanding of Ansible and Kubernetes terminologies is assumed, the steps are simple enough for beginners to follow along.
Prerequisites
Before proceeding, ensure the following prerequisites are in place:
- Provision Infrastructure: Set up computing resources, such as 3 servers, for the cluster.
- Install Dependencies: Install the following dependencies on your Ansible server (can be one of your nodes or a dedicated server {recommended} for ansible):
- Git
- Python3
- Pip3
- Ansible
Setting Up the Cluster
Follow these steps to set up your Kubernetes cluster with Kubespray:
Step 1: Set Up SSH Keys
Generate SSH keys on the Ansible node and copy the key to all your cluster nodes:
ssh-keygen # Go with the defaults
ssh-copy-id <user>@<node-IP>
Step 2: Download and Configure Kubespray
Download the Kubespray GitHub repository and checkout the latest version:
git clone git@github.com:Kubernetes-sigs/Kubespray.git
cd Kubespray
git checkout release-2.xx #replace 'xx' with release number
Step 3: Install Python Dependencies
Install the required Python dependencies using pip:
pip3 install -r ./requirements.txt
Step 4: Update Ansible Inventory
Update the Ansible inventory file with the IP addresses of your nodes:
cp -rfp inventory/sample inventory/mycluster
declare -a IPS=(<node1-IP> <node2-IP> <node3-IP>)
CONFIG_FILE=inventory/mycluster/hosts.yaml python3 contrib/inventory_builder/inventory.py ${IPS[@]}
Further customize inventory/mycluster/hosts.yaml to specify your master, worker, and etcd nodes.
Step 5: Review and Customize Configuration
Review and customize parameters under inventory/mycluster/group_vars:
cat inventory/mycluster/group_vars/all/all.yml
cat inventory/mycluster/group_vars/k8s_cluster/k8s-cluster.yml
Step 6: Allow Kubernetes Ports
If behind a firewall, ensure all necessary Kubernetes ports are allowed. These ports include; api-server, etcd, DNS, kubelet, kube-controller-manager, calico ports, kube-scheduler e.t.c.
Step 7: Clean Up Old Kubernetes Cluster
Run the playbook to clean up any old Kubernetes cluster:
ansible-playbook -i inventory/mycluster/hosts.yaml --user=<your-user-with-sudo-access> --ask-become-pass --become reset.yml
Step 8: Deploy Kubernetes with Kubespray
Run the playbook to deploy Kubernetes:
ansible-playbook -i inventory/my-cluster/hosts.yml --user=<your-user-with-sudo-access> --ask-become-pass --become cluster.yml
Step 9: Access the Cluster
To access the cluster using kubectl commands, update your kubeconfig file to be in the ~/.kube directory:
mkdir .kube
cd .kube/
sudo cp /etc/kubernetes/admin.conf config
kubectl get nodes
Note: The playbook will take some time to complete, but once finished, you'll have a highly available and self-managed Kubernetes cluster at your disposal.
Kubespray enables you to manage Kubernetes clusters, you can add/remove nodes (playbooks/scale.yml|remove_node.yml), delete the entire cluster (playbooks/reset.yml), and performing other Kubernetes management tasks.
Troubleshooting
- Issue with Python Packages Installation: When Ansible is already installed via system packages on the control node, Python packages installed using sudo pip install -r requirements.txt may end up in a different directory tree (e.g., /usr/local/lib/python2.7/dist-packages on Ubuntu) compared to Ansible's directory (e.g., /usr/lib/python2.7/dist-packages/ansible on Ubuntu). Consequently, the ansible-playbook command may fail with the following error:
ERROR! no action detected in task. This often indicates a misspelled module name, or incorrect module path.
This likely indicates that a task depends on a module present in requirements.txt.
2.Ensure Firewall Rules Allow Necessary Ports: Make sure that all necessary ports are allowed through the firewall to ensure proper communication between components.
3.Failure to Run Playbook without --become: The playbook will fail to run if the --become flag is not used. Ensure that you include --become to grant necessary privileges for the playbook to execute successfully.
4.For further troubleshooting on any encountered issue, please refer to the official Kubespray repository for comprehensive troubleshooting steps.
Top comments (1)
Great 👍