spiderpool-k8s-aws

Overview

The Spiderpool project aims to manage IP addresses in a Kubernetes cluster using Spiderpool and Multus CNI plugin. This project utilizes Terraform and Ansible to provision AWS EC2 instances and deploy k8s with spiderpool on them.

Installation

Prerequisites:

1. AWS account with necessary permissions and credentials stored in ~/.aws/credentials.
2. terraform (v0.12 or later).
3. ansible (v2.9 or later).
4. ssh public key in ~/.ssh/id_ed25519.pub.
5. helm.
6. kubectl.

Dependencies:

ansible-galaxy collection install kubernetes.core
pip install "Jinja2<3.1"
pip install kubernetes

Steps:

Clone the repo:

git clone https://github.com/pczarnik/spiderpool-k8s-aws.git
cd spiderpool-k8s-aws

Run provisioner and deployer script:

./provision_and_deploy.sh

This script will execute the necessary Terraform and Ansible commands to set up the environment, deploy Kubernetes, Spiderpool, and the applications.

Additional manual configuration

It's possible to add new worker in the variables.tf file. Ensure that SpiderIPPool.yaml file correctly defines new network configurations and IP pools.

Verify deployment

Access the Kubernetes cluster via kubectl and verify node and pod status. Check if nodes were correctly initalized, pods were created and spiderpool had correctly assigned IPs to them from the previously defined pools.

$ export KUBECONFIG="./kubeconfig"

$ kubectl get nodes
NAME      STATUS   ROLES           AGE   VERSION
master    Ready    control-plane   3m    v1.30.2
worker1   Ready    <none>          1m    v1.30.2
worker2   Ready    <none>          1m    v1.30.2
worker3   Ready    <none>          1m    v1.30.2
worker4   Ready    <none>          1m    v1.30.2

$ kubectl get spidermultusconfigs.spiderpool.spidernet.io -n kube-system
NAME          AGE
ipvlan-eth0   22m
ipvlan-eth1   22m

$ kubectl get spiderippools
NAME          VERSION   SUBNET           ALLOCATED-IP-COUNT   TOTAL-IP-COUNT
172-31-64-0   4         172.31.64.0/20   0                    16
172-31-96-0   4         172.31.96.0/20   0                    16

$ kubectl get pods -owide
NAME                       READY   STATUS    AGE   IP            NODE
busybox-5dc5cb4bf6-6kv5d   1/1     Running   24m   172.31.65.3   worker1
busybox-5dc5cb4bf6-9hxxb   1/1     Running   24m   172.31.66.2   worker2
busybox-5dc5cb4bf6-9jsts   1/1     Running   24m   172.31.68.1   worker4
busybox-5dc5cb4bf6-9ml79   1/1     Running   24m   172.31.67.4   worker3

$ kubectl exec -it busybox-5dc5cb4bf6-6kv5d -- ip -4 addr show scope global
2: eth0@eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9001 qdisc noqueue 
    inet 172.31.65.3/20 brd 172.31.79.255 scope global eth0
       valid_lft forever preferred_lft forever
4: net1@veth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9001 qdisc noqueue 
    inet 172.31.97.3/20 brd 172.31.111.255 scope global net1
       valid_lft forever preferred_lft forever