Prerequisites

  1. At least three VMs are required for K8s deployment.

  2. Servers required:

    a. Ansible Host

    b. Master Node

    c. Worker Nodes

  3. Ansible should be installed on Ansible Host. For installation steps, follow install-ansible.md.

How to use this (Setup Instructions)

  1. Make servers ready (Ansible Host, one master node and multiple worker nodes) by creating new VMs on Fiware-Lab.
  2. Update the Kernel if it is lower than 5.x (for CentOS). Steps are added in update-kernel.md.
  3. Install firewall (if not installed). Steps are added in setup-firewalld.md.

  4. Make an entry of each hosts in /etc/hosts file for name resolution in Master and Worker VM as:

    On Master: 127.0.0.1 <master-vm-name>.novalocal

    On Worker: 127.0.0.1 <worker-vm-name>.novalocal

  5. Make sure kubernetes master node and other worker nodes are reachable between each other.

  6. Internet connection must be enabled in all nodes, required packages will be downloaded from kubernetes official yum repository.
  7. Copy the script and ansible playbooks on Ansible Host VM.
  8. There is a file hosts in this folder, make entries of all kubernetes nodes.
  9. Provide server details in playbooks/env_variables.
  10. Deploy the ssh key from Ansible Host to nodes for password less authentication. Steps are added in passwordless-ssh.md.
  11. Provide root level privilege to centos or ubuntu user. Steps are added in root-access.md.

  12. Run settingup_kubernetes_cluster.yml playbook to setup all nodes and kubernetes master configuration.

    ansible-playbook settingup_kubernetes_cluster.yml

  13. Run join_kubernetes_workers_nodes.yml playbook to join the worker nodes with kubernetes master node once settingup_kubernetes_cluster.yml playbook tasks are completed.

    ansible-playbook join_kubernetes_workers_nodes.yml

  14. Verify the configuration from master node:

    kubectl get nodes kubectl get pods --all-namespaces

Troubleshooting

In case settingup_kubernetes_cluster playbook fails due to some system errors. Then reset the cluster otherwise it will show different errors like PORT IN USE and CERTIFICATES ALREADY EXIST To fix these errors run below commands on master node:

kubeadm reset
iptables -F && iptables -t nat -F && iptables -t mangle -F && iptables -X

Deploy Fiware GEs using HELM

  1. To make use of the charts, you may add the repository:

    helm repo add fiware https://fiware.github.io/helm-charts 2. After the repo is added all charts can be installed via:

    helm install <RELEASE_NAME> fiware/<CHART_NAME>

Limitations

  1. Ansible script would deploy bare minimum K8s. User can configure K8s like persistent storage etc.
  2. At least three VMs are required for K8s deployment.