KubeEdge-Openstack-Ansible-Automation

This is the repository for automatically deploying the KubeEdge to OpenStack based cloud using ansible and terraform.

Deployment Architecture

Overall deployment automation architecture of IoT platform:

Requirements

The following components are necessary for setting up and interacting with the cluster:

• KubeEdge
• terraform
• ansible
• docker

Currently following are enabled/deployment as part of this automation

• VMs creation (1 Master, 1 Worker, 1 Edge; can be changed in variables.tf)
• Docker
• Kubernetes usig Kubeadm
• Flannel used as CNI
• Helm
• Keadm v1.9.1 both on Master and Edge
• Keadm logs enabled both on Master and Edge
• Edgemesh both on Master and Edge (https://edgemesh.netlify.app/guide/getting-started.html#dependencies)
• Monitoring: Prometheus, Grafana, Kubernetes Metrics Server for edge
• OpenFaaS
• Sedna

Get Started

1. Clone the repository and go to the root directory:

git https://github.com/ansjin/KubeEdge-Openstack-Ansible-Automation
cd KubeEdge-Openstack-Ansible-Automation

2. Create a key-pair in the cloud

TODO

3. Replace the key variables in the config files accordingly

variables.tf (instance_keypair_name)
variables.tf (ssh_key_file)
ansible/ansible.cfg (private_key_file)
# all below should be equivalent to openstack_user_name
terraform.tfvars (openstack_user_name)
terraform.tfvars (openstack_tenant_name)
terraform.tfvars (openstack_project_name)
terraform.tfvars (openstack_project_id)

4. Download the credentials from the cloud and source it

source xxxxxxx-openrc.sh

5. Initialize the terraform providers

terraform init

6. Bootstrap the cluster

terraform apply -auto-approve

7. Destroy the cluster

terraform destroy -auto-approve

Run Ansible Commands Separately

cd ansible
ansible-playbook master.yml -vvvv
ansible-playbook worker.yml -vvvv
ansible-playbook master-after-join.yml -vvvv
# ansible-playbook -i cloud_hosts.ini --start-at-task "Install K8s dashboard" master.yml -vvvv

Dashboard Links

• Kubernetes Dashboard (http://MASTER_PUBLIC_IP:32007)

• Grafana Dashboard (http://MASTER_PUBLIC_IP:32006)

• Prometheus Configuration for Grafana (http://MASTER_PUBLIC_IP:32009)

NOTE: It takes about 10 min for Terraform to run + 5 min for all the Pods to be Running/Completed.