clinco - Cluster in Container

---

BEFORE WE START

Let us take a look at the Docker images used for nodes

The ubuntu based images "petschenek/ubuntu-systemd:master-arc-21.10 (compressed size 220MB) and petschenek/ubuntu-systemd:worker-arc-21.10 (compressed size 230MB)" have systemd installed in them. You can check the Dockerfile for master-amd64 and Dockerfile for worker-amd64

In these Dockerfiles there are three important points:

1. We need to tell systemd that it is in Docker. ENV container=docker
2. We need to declare the stopsignal different than the default stopsignal of Docker which is SIGTERM whereas systemd exits on SIGRTMIN+3. STOPSIGNAL SIGRTMIN+3
3. We need to declare a volume for /var/lib/containerd in worker node image because the filesystem of pods that will be created in the cluster must be a normal filesystem. When you run Docker in Docker, the outer Docker runs on top of a normal filesystem (EXT4, BTRFS etc), but the inner Docker runs on top of a copy-on-write system (AUFS, BTRFS etc). Since you cannot run AUFS on top of AUFS, you need to use VOLUME ["/var/lib/containerd"] in the Dockerfile or specify -v /var/lib/containerd during container creation.

---

PREREQUISITES

• DOCKER

• KUBECTL

• HELM

Must be installed on your host

---

LET'S START

1) Create a cluster

Download the scripts:

• git clone --depth=1 https://github.com/ManasPecenek/clinco.git && cd clinco && export KUBECONFIG=./admin.kubeconfig && alias startCluster="bash $PWD/initial-script.sh" addNode="bash $PWD/add-worker.sh"

Now run the script with how many worker nodes you want. For example "startCluster -n 3" will result in a 3-worker-node cluster:

• startCluster -n <worker-node-count>

If you do not specify worker node count, it will be "1" by default.

2) Check the cluster via Installing OPA Gatekeeper

• docker ps

• kubectl cluster-info Note: If you cannot connect to the cluster by default, just run "export KUBECONFIG=./admin.kubeconfig" in the same directory.

• kubectl get nodes -o wide

• kubectl apply -f https://raw.githubusercontent.com/open-policy-agent/gatekeeper/release-3.7/deploy/gatekeeper.yaml

• kubectl get pods -o wide -n gatekeeper-system

3) Add additional worker nodes

If you want to add additional worker nodes, all you need to do is run addNode -n <number>. For example addNode -n 2 will add two additional nodes into your cluster. If you do not specify a number and run only addNode, it will add only 1 node to your cluster.

---

RESOURCES

• Kubernetes the Hard Way
• Nginx Ingress Controller
• Local Storage Class