Getting Started

To play around I suggest using Skate in Docker (SIND) to run a local cluster. You can use ./hack/sindplz to create a cluster of 2 nodes.

Shell

# assumes an ssh priv pub key combo ~/.ssh/id_rsa & ~/.ssh/id_rsa.pub that it will add to the container host's authorized_keys
# create a hack/.sindplz.env file with SSH_PRIVATE_KEY and SSH_PUBLIC_KEY set to override this
./hack/sindplz create

BTW: you can run ./hack/sindplz create again to run a new cluster if things get messed up.

Download the skate binary for your platform and architecture.

# Get list of latest release binaries
curl -s https://api.github.com/repos/skateco/skate/releases/latest | grep "browser_download_url.*tar.gz" | cut -d : -f 2,3 | tr -d \\\" | tr -d "[:blank:]"|grep -v skatelet

Put it in your path.

# feel free to put it whereever but if you don't care...
sudo mv skate /usr/local/bin

Now, let’s register a cluster:

Note: Change ~/.ssh/id_rsa to the path to the private key that can access your nodes. Should be the same that was used with the sindplz script

skate create cluster my-cluster --default-user $USER --default-key ~/.ssh/id_rsa

Add the nodes:

# automatically terraform the 2 nodes
> ./hack/sindplz skate

Ok, now we should have a 2 node cluster that we can deploy to.

# list the nodes to be sure
> skate get nodes
NAME                            PODS        STATUS    
node-1                          2           Healthy   
node-2                          2           Healthy

Create a deployment

# creates an nginx deployment with 2 replicas
cat <<EOF | skate apply -f -
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx
  namespace: my-app
spec:
  replicas: 2
  template:
    spec:
      containers:
      - name: nginx
        image: nginx:1.14.2
EOF

Check the deployment

skate get deployment

NAMESPACE  NAME   READY  UPTODATE  AVAILABLE  AGE 
my-app     nginx  2/2    2         2          45s

Now you can create a service:

cat <<EOF | skate apply -f -
---
apiVersion: v1
kind: Service
metadata:
  name: nginx
  namespace: my-app
spec:
  selector:
    app.kubernetes.io/name: nginx
  ports:
  - protocol: TCP
    port: 80
    targetPort: 80
EOF

skate get service

NAMESPACE  NAME   CLUSTERIP  EXTERNALIP  PORTS  AGE 
my-app     nginx  -          -           80     10s

And finally an ingress:

cat <<EOF | skate apply -f -
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: public
  namespace: my-app
  annotations:
    # since the vm cluster isn't public, letsencrypt wont work, so turn off redirect so we can test
    nginx.ingress.kubernetes.io/ssl-redirect: "false"
spec:
  rules:
  - host: nginx.example.com
    http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: nginx.my-app
            port:
              number: 80
EOF

Now let’s do a quick request against the cluster:

# 192.168.76.11 is a node-ip from hack/sindplz ips
curl --header "Host: nginx.example.com" --insecure  http://192.168.76.11

You should see something like this:

<!DOCTYPE html>
<html>
<head>
<title>Welcome to nginx!</title>
...

Great Success!!

Now you’ve deployed a webservice across 2 nodes available via the ingress.

Skate Components