New to Voyager? Please start here.

Specify NodePort

If you are using a NodePort or LoadBalancer type Ingress, a NodePort or LoadBalancer type Service is used to expose HAProxy pods respectively. If no node port is specified for each HAProxy Service port, Kubernetes will randomly assign one for you.

Since 3.2.0, you have the option to specify a NodePort for each HAProxy Service port. This allows you to guarantee that the port will not get changed, as you make changes to an Ingress object. If you specify nothing, Kubernetes will auto assign as before.

Ingress Example

First create a test-server and expose it via service:

$ kubectl run test-server --image=gcr.io/google_containers/echoserver:1.8
deployment "test-server" created

$ kubectl expose deployment test-server --type=LoadBalancer --port=80 --target-port=8080
service "test-server" exposed

Then create the ingress:

apiVersion: voyager.appscode.com/v1
kind: Ingress
metadata:
  name: test-ingress
  namespace: default
  annotations:
    ingress.appscode.com/type: NodePort
spec:
  rules:
  - host: one.example.com
    http:
      port: 8989
      nodePort: 32666
      paths:
      - path: /t1
        backend:
          service:
            name: test-server
            port:
              number: 80
      - path: /t2
        backend:
          service:
            name: test-server
            port:
              number: 80
  - host: other.example.com
    http:
      port: 8989
      nodePort: 32666
      paths:
      - backend:
          service:
            name: test-server
            port:
              number: 80
  - host: appscode.example.com
    tcp:
      port: 4343
      nodePort: 32667
      backend:
        service:
          name: test-server
          port:
            number: 80

Since ingress.appscode.com/type: NodePort annotation is used, this Ingress is going to expose HAProxy pods via a NodePort Service. This service will listen to 8989 and 4343 port for incoming HTTP connections and these port will map to specified node ports, and will pass any request coming to it to the desired backend.

$ kubectl get pods,svc
NAME                                      READY     STATUS    RESTARTS   AGE
po/test-server-68ddc845cd-x7dtv           1/1       Running   0          23h
po/voyager-test-ingress-77cc5d54d-sgzkv   1/1       Running   0          18s

NAME                       TYPE           CLUSTER-IP      EXTERNAL-IP   PORT(S)                         AGE
svc/kubernetes             ClusterIP      10.96.0.1       <none>        443/TCP                         2d
svc/test-server            LoadBalancer   10.105.13.31    <pending>     80:30390/TCP                    1d
svc/voyager-test-ingress   NodePort       10.106.53.141   <none>        8989:32666/TCP,4343:32667/TCP   26m
$ kubectl get svc voyager-test-ingress -o yaml

apiVersion: v1
kind: Service
metadata:
  annotations:
    ingress.appscode.com/last-applied-annotation-keys: ""
    ingress.appscode.com/origin-api-schema: voyager.appscode.com/v1
    ingress.appscode.com/origin-name: test-ingress
  creationTimestamp: 2018-02-15T03:51:06Z
  name: voyager-test-ingress
  namespace: default
  ownerReferences:
  - apiVersion: voyager.appscode.com/v1
    blockOwnerDeletion: true
    kind: Ingress
    name: test-ingress
    uid: 73203752-1203-11e8-b2d5-080027eaa7b2
  resourceVersion: "65769"
  selfLink: /api/v1/namespaces/default/services/voyager-test-ingress
  uid: 732a5322-1203-11e8-b2d5-080027eaa7b2
spec:
  clusterIP: 10.106.53.141
  externalTrafficPolicy: Cluster
  ports:
  - name: tcp-8989
    nodePort: 32666
    port: 8989
    protocol: TCP
    targetPort: 8989
  - name: tcp-4343
    nodePort: 32667
    port: 4343
    protocol: TCP
    targetPort: 4343
  selector:
    origin: voyager
    origin-api-group: voyager.appscode.com
    origin-name: test-ingress
  sessionAffinity: None
  type: NodePort
status:
  loadBalancer: {}

Now, if you check the HAProxy configuration generated by Voyager, you should see something like below:

# HAProxy configuration generated by https://github.com/appscode/voyager
# DO NOT EDIT!
global
	daemon
	stats socket /tmp/haproxy
	server-state-file global
	server-state-base /var/state/haproxy/
	# log using a syslog socket
	log /dev/log local0 info
	log /dev/log local0 notice
	tune.ssl.default-dh-param 2048
	ssl-default-bind-ciphers ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-AES256-GCM-SHA384:DHE-RSA-AES128-GCM-SHA256:DHE-DSS-AES128-GCM-SHA256:kEDH+AESGCM:ECDHE-RSA-AES128-SHA256:ECDHE-ECDSA-AES128-SHA256:ECDHE-RSA-AES128-SHA:ECDHE-ECDSA-AES128-SHA:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES256-SHA384:ECDHE-RSA-AES256-SHA:ECDHE-ECDSA-AES256-SHA:DHE-RSA-AES128-SHA256:DHE-RSA-AES128-SHA:DHE-DSS-AES128-SHA256:DHE-RSA-AES256-SHA256:DHE-DSS-AES256-SHA:DHE-RSA-AES256-SHA:!aNULL:!eNULL:!EXPORT:!DES:!RC4:!3DES:!MD5:!PSK
	hard-stop-after 30s
defaults
	log global
	# https://cbonte.github.io/haproxy-dconv/1.7/configuration.html#4.2-option%20abortonclose
	# https://github.com/voyagermesh/voyager/pull/403
	option dontlognull
	option http-server-close
	# Timeout values
	timeout client 50s
	timeout client-fin 50s
	timeout connect 5s
	timeout server 50s
	timeout tunnel 50s
	# Configure error files
	# default traffic mode is http
	# mode is overwritten in case of tcp services
	mode http
frontend http-0_0_0_0-8989
	bind *:8989
	mode http
	option httplog
	option forwardfor
	acl is_proxy_https hdr(X-Forwarded-Proto) https
	acl acl_other.example.com hdr(host) -i other.example.com:8989
	use_backend test-server.default:80-2bdf8f33305898e39d66486c50d39fc1 if acl_other.example.com
	acl acl_one.example.com hdr(host) -i one.example.com:8989
	acl acl_one.example.com:t2 path_beg /t2
	use_backend test-server.default:80-64e63a31b2e805238363fc7982c38f12 if acl_one.example.com acl_one.example.com:t2
	acl acl_one.example.com:t1 path_beg /t1
	use_backend test-server.default:80-6c5cadcbfcb85a324f0cf5c4654dd952 if acl_one.example.com acl_one.example.com:t1
backend test-server.default:80-2bdf8f33305898e39d66486c50d39fc1
	server pod-test-server-68ddc845cd-x7dtv 172.17.0.4:8080
backend test-server.default:80-64e63a31b2e805238363fc7982c38f12
	server pod-test-server-68ddc845cd-x7dtv 172.17.0.4:8080
backend test-server.default:80-6c5cadcbfcb85a324f0cf5c4654dd952
	server pod-test-server-68ddc845cd-x7dtv 172.17.0.4:8080
frontend tcp-0_0_0_0-4343
	bind *:4343
	mode tcp
	default_backend test-server.default:80-b700282ee9f2823f5b4ad3452658a791
backend test-server.default:80-b700282ee9f2823f5b4ad3452658a791
	mode tcp
	server pod-test-server-68ddc845cd-x7dtv 172.17.0.4:8080

Port 8989 has 2 separate hosts one.example.com and other.example.com. one.example.com has 2 paths /t1 and /t2. Since they all are exposed via the same HTTP port, they must use the same NodePort.