ome / kubernetes-platform

Infrastructure for managing the Kubernetes platform in the OME

Infrastructure for managing the Kubernetes platform in the OME

WARNING: This repository is in development and is not ready for a production deployment of Kubernetes

This repository contains Ansible files for deploying a multi-master high-availability Kubernetes cluster, and Terraform configuration for provisioning virtual machines on openstack.

The majority of the work is handled by the KubeSpray project. A fork exists in the openmicroscopy organization.

Example usage

This example uses a configuration defined under kubetest. You can either modify this configuration, or copy the directory.

Create some virtual machines on openstack using the configuration in kubetest/kubespray.tf and kubetest/kubetest.tfvars:

ENV=kubetest
./run-terraform.sh $ENV init
./run-terraform.sh $ENV get
./run-terraform.sh $ENV apply

Terraform uses a state file $ENV/terraform.tfstate and does not take existing infrastructure into account, so if multiple users are maintaining the same deployment this file must be shared. These terraform commands are idempotent as long as this tfstate file is present. If you run terraform without this state file it will create a new set of infrastructure (with the same names).

Install Kubernetes using a dynamic terraform inventory based on the tfstate file, and inventory variables in kubetest/inventory/inventory.yml:

./run-ansible.sh $ENV

If this works you should have a kubernetes admin file in kubespray/artifacts/admin.conf. Copy this, and edit the clusters.cluster.server key to point to the external address of the Kubernetes cluster.

Connect to the cluster using a local kubectl client:

$ kubectl --kubeconfig=admin.conf get nodes

NAME                     STATUS    ROLES     AGE       VERSION
kubetest-k8s-master-1    Ready     master    5d        v1.8.3+coreos.0
kubetest-k8s-node-nf-1   Ready     node      5d        v1.8.3+coreos.0
kubetest-k8s-node-nf-2   Ready     node      5d        v1.8.3+coreos.0