You are looking at the documentation of a prior release. To read the documentation of the latest release, please visit here.

LoadBalancer

In LoadBalancer type Ingress, HAProxy pods are exposed via a LoadBalancer type Kubernetes service named voyager-${ingress-name}. You can apply the ingress.appscode.com/type: LoadBalancer annotation on a Ingress object to enable this type of Ingress. This is also the default type for Ingress objects. So, this annotaion is not required to enable this type.

How It Works

  • First, deploy voyager operator.
curl -fsSL https://raw.githubusercontent.com/appscode/voyager/6.0.0/hack/deploy/voyager.sh \
    | bash -s -- --provider=gke
curl -fsSL https://raw.githubusercontent.com/appscode/voyager/6.0.0/docs/examples/ingress/types/loadbalancer/deploy-servers.sh | bash

deployment "nginx" created
service "web" exposed
deployment "echoserver" created
service "rest" exposed
  • Now, create an Ingress object running kubectl apply -f https://raw.githubusercontent.com/appscode/voyager/6.0.0/docs/examples/ingress/types/loadbalancer/ing.yaml. Please note the annotaiton on ingress:
  annotations:
    ingress.appscode.com/type: LoadBalancer
$ kubectl get pods,svc
NAME                                       READY     STATUS    RESTARTS   AGE
po/echoserver-848b75d85-wxdrz              1/1       Running   0          2m
po/nginx-7c87f569d-5q5mf                   1/1       Running   0          3m
po/voyager-test-ingress-687d6b7c65-qjqzt   1/1       Running   0          1m

NAME                       TYPE           CLUSTER-IP      EXTERNAL-IP      PORT(S)        AGE
svc/kubernetes             ClusterIP      10.11.240.1     <none>           443/TCP        4m
svc/rest                   ClusterIP      10.11.252.242   <none>           80/TCP         2m
svc/voyager-test-ingress   LoadBalancer   10.11.248.185   35.226.114.148   80:30854/TCP   1m
svc/web                    ClusterIP      10.11.253.33    <none>           80/TCP         2m
$ curl -vv 35.226.114.148 -H "Host: web.example.com"
* Rebuilt URL to: 35.226.114.148/
*   Trying 35.226.114.148...
* Connected to 35.226.114.148 (35.226.114.148) port 80 (#0)
> GET / HTTP/1.1
> Host: web.example.com
> User-Agent: curl/7.47.0
> Accept: */*
>
< HTTP/1.1 200 OK
< Server: nginx/1.13.8
< Date: Thu, 18 Jan 2018 06:40:49 GMT
< Content-Type: text/html
< Content-Length: 612
< Last-Modified: Tue, 26 Dec 2017 11:11:22 GMT
< ETag: "5a422e5a-264"
< Accept-Ranges: bytes
<
<!DOCTYPE html>
<html>
<head>
<title>Welcome to nginx!</title>
<style>
    body {
        width: 35em;
        margin: 0 auto;
        font-family: Tahoma, Verdana, Arial, sans-serif;
    }
</style>
</head>
<body>
<h1>Welcome to nginx!</h1>
<p>If you see this page, the nginx web server is successfully installed and
working. Further configuration is required.</p>

<p>For online documentation and support please refer to
<a href="http://nginx.org/">nginx.org</a>.<br/>
Commercial support is available at
<a href="http://nginx.com/">nginx.com</a>.</p>

<p><em>Thank you for using nginx.</em></p>
</body>
</html>
* Connection #0 to host 35.226.114.148 left intact
$ curl -vv 35.226.114.148 -H "Host: app.example.com"
* Rebuilt URL to: 35.226.114.148/
*   Trying 35.226.114.148...
* Connected to 35.226.114.148 (35.226.114.148) port 80 (#0)
> GET / HTTP/1.1
> Host: app.example.com
> User-Agent: curl/7.47.0
> Accept: */*
>
< HTTP/1.1 200 OK
< Server: nginx/1.10.0
< Date: Thu, 18 Jan 2018 06:41:36 GMT
< Content-Type: text/plain
< Transfer-Encoding: chunked
<
CLIENT VALUES:
client_address=10.8.0.14
command=GET
real path=/
query=nil
request_version=1.1
request_uri=http://app.example.com:8080/

SERVER VALUES:
server_version=nginx: 1.10.0 - lua: 10001

HEADERS RECEIVED:
accept=*/*
connection=close
host=app.example.com
user-agent=curl/7.47.0
x-forwarded-for=10.8.0.1
BODY:
* Connection #0 to host 35.226.114.148 left intact

FAQ

How do I ensure that IP assigned my Ingress does not change?

You can allocate a static IP to a LoadBalancer Ingress managed by Voyager. Say for example, you are using GKE. When you create an Ingress object, Voyager will create a Kubernetes Service of type LoadBalancer. This service will will automatically get a regional IP. If you want to keep that IP, you can mark that IP as static in Google cloud console and the apply the annotation to your Ingress.

  annotations:
    ingress.appscode.com/type: LoadBalancer
 `  ingress.appscode.com/load-balancer-ip: 'a.b.c.d'`

How do I use a Global Static IP (anycast IP) with an Ingress in GKE?

You can’t use Global Static IP with a LoabBalancer Ingress managed by GKE. Voyager creates a LoadBalancer Service to expose HAProxy pods. Under the hood, Kubernetes creates a Network LoadBalancer to expose that Kubernetes service. Network LoadBalancers can only use regional static IPs.

If you want to use Global static IP with Google Cloud, these pods need to be exposed via a HTTP LoadBalancer. Voyager does not support this today. This is not a priority for us but if you want to contribute, we can talk more. To use HTTP LoadBalancers today, you can use the gce ingress controller: https://github.com/kubernetes/ingress-gce . You may already know that HTTP LoadBalancer can only open port 80, 8080 and 443 and serve HTTP traffic. Please consult the official docs for more details: https://cloud.google.com/compute/docs/load-balancing/

How to use LoadBalancer type ingress in Openstack?

If you need to create an internal LB in Openstack, you can do so using ingress.appscode.com/annotations-service annotation on the Ingress object.

  annotations:
    ingress.appscode.com/type: LoadBalancer
    ingress.appscode.com/annotations-service: |
      {
        "service.beta.kubernetes.io/openstack-internal-load-balancer": "true"
      }      

How to use LoadBalancer type ingress in Minikube cluster?

Minikube clusters do not support service type LoadBalancer. So, you can try the following work arounds:

  • You can set the Host header is your http request to match the expected domain and port. This will ensure HAProxy matches the rules properly.
$ curl -vv <minikube-ip>:<node-port> -H "Host: app.example.com"
  • This work around is available thanks to @david92rl. You can use a service type ClusterIP with an ip fixed (like 10.0.0.150), then create a route to it from host machine.

Minikube on Mac with virtualbox/vmware providers

sudo route -n delete ${K8S_NETWORK} > /dev/null 2>&1
sudo route -n add ${K8S_NETWORK} $(minikube ip)
interface=$(ifconfig 'bridge0' | grep member | awk '{print $2}' | xargs | awk '{print $1}')
sudo ifconfig bridge0 -hostfilter ${interface}

Minikube on Linux

sudo ip route del ${K8S_NETWORK}
sudo ip route add ${K8S_NETWORK} via $(minikube ip)

K8S_NETWORK usually is 10.0.0.0/24 but it’s worth to double check always.