Deploying HAProxy Ingress Controller

Don't have a Kubernetes cluster? Single-node of CoreOS Kubernetes is a good starting point.

Deploy a default backend used to serve 404 Not Found pages:

kubectl run ingress-default-backend \
  --image=gcr.io/google_containers/defaultbackend:1.0 \
  --port=8080 \
  --limits=cpu=10m,memory=20Mi \
  --expose

Check if the default backend is up and running:

kubectl get pod
NAME                                       READY     STATUS    RESTARTS   AGE
ingress-default-backend-1110790216-gqr61   1/1       Running   0          10s

Deploy certificate and private key used to serve https on ingress that doesn't provide it's own certificate. For testing purposes a self signed certificate is ok:

openssl req \
  -x509 -newkey rsa:2048 -nodes -days 365 \
  -keyout tls.key -out tls.crt -subj '/CN=localhost'
kubectl create secret tls ingress-default-ssl --cert=tls.crt --key=tls.key
rm -v tls.crt tls.key

Deploy HAProxy Ingress. Note that hostNetwork: true could be uncommented if your cluster has IPs that doesn't use ports 80, 443 and 1936.

kubectl create -f haproxy-ingress.yaml

Check if the controller was successfully deployed:

kubectl get pod -w
NAME                                       READY     STATUS    RESTARTS   AGE
haproxy-ingress-2556761959-tv20k           1/1       Running   0          12s
ingress-default-backend-1110790216-gqr61   1/1       Running   0          3m
^C

Problem? Check logs and events of the POD:

kubectl logs haproxy-ingress-2556761959-tv20k
kubectl describe haproxy-ingress-2556761959-tv20k

Deploy some web application and it's ingress resource:

kubectl run nginx --image=nginx:alpine --port=80 --expose
kubectl create -f - <<EOF
  apiVersion: extensions/v1beta1
  kind: Ingress
  metadata:
    name: app
  spec:
    rules:
    - host: foo.bar
      http:
        paths:
        - path: /
          backend:
            serviceName: nginx
            servicePort: 80
EOF

Exposing HAProxy Ingress depend on your Kubernetes environment. If hostNetwork was defined just use host's public IP, otherwise expose the controller as a type=NodePort service:

kubectl expose deploy/haproxy-ingress --type=NodePort
kubectl get svc/haproxy-ingress -oyaml

Look for nodePort field next to port: 80.

Change below 172.17.4.99 to the host's IP and 30876 to the nodePort, or remove :30876 if using hostNetwork:

curl -i 172.17.4.99:30876
HTTP/1.1 404 Not Found
Date: Mon, 05 Feb 2017 22:59:36 GMT
Content-Length: 21
Content-Type: text/plain; charset=utf-8

default backend - 404

Using default backend because host was not found.

Now try to send a header:

curl -i 172.17.4.99:30876 -H 'Host: foo.bar'
HTTP/1.1 200 OK
Server: nginx/1.11.9
Date: Mon, 05 Feb 2017 23:00:33 GMT
Content-Type: text/html
Content-Length: 612
Last-Modified: Tue, 24 Jan 2017 18:53:46 GMT
ETag: "5887a2ba-264"
Accept-Ranges: bytes

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

Not what you were looking for? Have a look at controller's logs:

kubectl get pod
NAME                                       READY     STATUS    RESTARTS   AGE
haproxy-ingress-2556761959-tv20k           1/1       Running   0          9m
...

kubectl logs haproxy-ingress-2556761959-tv20k | less -S

3.5 KiB Raw Blame History

Deploying HAProxy Ingress Controller

3.5 KiB

Raw Blame History