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

New to Voyager? Please start here.

Installation Guide

Voyager operator can be installed via a script or as a Helm chart.

Using Script

To install Voyager in your Kubernetes cluster, pick the appropriate cluster provider and run the following command:

# provider=acs
# provider=aks
# provider=aws
# provider=azure
# provider=baremetal
# provider=gce
# provider=gke
# provider=minikube
# provider=openstack
# provider=metallb
# provider=digitalocean
# provider=linode

$ curl -fsSL https://raw.githubusercontent.com/appscode/voyager/9.0.0/hack/deploy/voyager.sh \
    | bash -s -- --provider=$provider

After successful installation, you should have a voyager-operator-*** pod running in the kube-system namespace.

$ kubectl get pods -n kube-system | grep voyager-operator
voyager-operator-846d47f489-jrb58       1/1       Running   0          48s

Customizing Installer

The installer script and associated yaml files can be found in the /hack/deploy folder. To see the full list of flags available to installer, use the -h flag.

$ curl -fsSL https://raw.githubusercontent.com/appscode/voyager/9.0.0/hack/deploy/voyager.sh | bash -s -- -h
voyager.sh - install voyager operator

voyager.sh [options]

options:
-h, --help                             show brief help
-n, --namespace=NAMESPACE              specify namespace (default: kube-system)
-p, --provider=PROVIDER                specify a cloud provider
    --rbac                             create RBAC roles and bindings (default: true)
    --docker-registry                  docker registry used to pull voyager images (default: appscode)
    --haproxy-image-tag                tag of Docker image containing HAProxy binary (default: 1.9.2-9.0.0-alpine)
    --image-pull-secret                name of secret used to pull voyager operator images
    --restrict-to-namespace            restrict voyager to its own namespace
    --run-on-master                    run voyager operator on master
    --enable-validating-webhook        enable/disable validating webhooks for voyager CRDs
    --bypass-validating-webhook-xray   if true, bypasses validating webhook xray checks
    --template-cfgmap=CONFIGMAP        name of configmap with custom templates
    --enable-status-subresource        if enabled, uses status sub resource for Voyager crds
    --use-kubeapiserver-fqdn-for-aks   if true, uses kube-apiserver FQDN for AKS cluster to workaround https://github.com/Azure/AKS/issues/522 (default true)
    --enable-analytics                 send usage events to Google Analytics (default: true)
    --uninstall                        uninstall voyager
    --purge                            purges Voyager crd objects and crds

If you would like to run Voyager operator pod in master instances, pass the --run-on-master flag:

$ curl -fsSL https://raw.githubusercontent.com/appscode/voyager/9.0.0/hack/deploy/voyager.sh \
    | bash -s -- --provider=$provider --run-on-master [--rbac]

Voyager operator will be installed in a kube-system namespace by default. If you would like to run Voyager operator pod in voyager namespace, pass the --namespace=voyager flag:

$ kubectl create namespace voyager
$ curl -fsSL https://raw.githubusercontent.com/appscode/voyager/9.0.0/hack/deploy/voyager.sh \
    | bash -s -- --provider=$provider --namespace=voyager [--run-on-master] [--rbac]

By default, Voyager operator will watch Ingress objects in any namespace. If you would like to restrict Voyager to Ingress and Services in its own namespace, pass the --restrict-to-namespace flag:

$ kubectl create namespace voyager
$ curl -fsSL https://raw.githubusercontent.com/appscode/voyager/9.0.0/hack/deploy/voyager.sh \
    | bash -s -- --provider=$provider --namespace=voyager --restrict-to-namespace [--run-on-master] [--rbac]

If you are using a private Docker registry, you need to pull the following 2 docker images:

To pass the address of your private registry and optionally a image pull secret use flags --docker-registry and --image-pull-secret respectively.

$ kubectl create namespace voyager
$ curl -fsSL https://raw.githubusercontent.com/appscode/voyager/9.0.0/hack/deploy/voyager.sh \
    | bash -s -- --provider=$provider --docker-registry=MY_REGISTRY [--image-pull-secret=SECRET_NAME] [--rbac]

By default, Voyager uses Alpine based HAProxy image (1.9.2-9.0.0-alpine). But you can also Debian based image for HAProxy by setting –haproxy-image-tag=1.9.2-9.0.0 flag.

Voyager implements a validating admission webhook to validate Voyager CRDs. This is enabled by default for Kubernetes 1.9.0 or later releases. To disable this feature, pass the --enable-validating-webhook=false flag.

$ curl -fsSL https://raw.githubusercontent.com/appscode/voyager/9.0.0/hack/deploy/voyager.sh \
    | bash -s -- --provider=$provider --enable-validating-webhook [--rbac]

Voyager 9.0.0 or later releases can use status sub resource for CustomResourceDefintions. This is enabled by default for Kubernetes 1.11.0 or later releases. To disable this feature, pass the --enable-status-subresource=false flag.

To use custom templates to render HAProxy configuration, visit here.

Using Helm

Voyager can be installed via Helm 2.9.x or later versions using the chart from AppsCode Charts Repository. To install the chart with the release name my-release:

$ helm repo add appscode https://charts.appscode.com/stable/
$ helm repo update
$ helm search appscode/voyager
NAME              CHART VERSION APP VERSION DESCRIPTION
appscode/voyager  9.0.0    9.0.0  Voyager by AppsCode - Secure HAProxy Ingress Controller...

# provider=acs
# provider=aks
# provider=aws
# provider=azure
# provider=baremetal
# provider=gce
# provider=gke
# provider=minikube
# provider=openstack
# provider=metallb
# provider=digitalocean
# provider=linode

$ helm install appscode/voyager --name voyager-operator --version 9.0.0 \
  --namespace kube-system \
  --set cloudProvider=$provider

To see the detailed configuration options, visit here.

Installing in GKE Cluster

If you are installing Voyager on a GKE cluster, you will need cluster admin permissions to install Voyager operator. Run the following command to grant admin permision to the cluster.

$ kubectl create clusterrolebinding "cluster-admin-$(whoami)" \
  --clusterrole=cluster-admin \
  --user="$(gcloud config get-value core/account)"

Installing in Minikube

Voyager can be used in minikube using --provider=minikube. In Minikube, a LoadBalancer type ingress will only assigned a NodePort.

Installing in Baremetal Cluster

Voyager works great in baremetal cluster. To install, set --provider=baremetal. In baremetal cluster, LoadBalancer type ingress in not supported. You can use NodePort, HostPort or Internal ingress objects.

Installing in Baremetal Cluster with MetalLB

Follow the instructions for installing on baremetal cluster but specify metallb as provider. Then install MetalLB following the instructions here. Now, you can use LoadBalancer type ingress in baremetal clusters.

Installing in DigitalOcean Cluster

To use LoadBalancer type ingress in DigitalOcean cluster, install Kubernetes cloud controller manager for DigitalOcean. Otherwise set cloud provider to barematal.

Installing in Linode Cluster

To use LoadBalancer type ingress in Linode cluster, install Kubernetes cloud controller manager for Linode. Otherwise set cloud provider to barematal.

Verify installation

To check if Voyager operator pods have started, run the following command:

$ kubectl get pods --all-namespaces -l app=voyager --watch

Once the operator pods are running, you can cancel the above command by typing Ctrl+C.

Now, to confirm CRD groups have been registered by the operator, run the following command:

$ kubectl get crd -l app=voyager

Now, you are ready to create your first ingress using Voyager.

Configuring RBAC

Voyager creates two CRDs: Ingress and Certificate. Voyager installer will create 2 user facing cluster roles:

ClusterRoleAggregates ToDesription
appscode:voyager:editadmin, editAllows edit access to Voyager CRDs, intended to be granted within a namespace using a RoleBinding.
appscode:voyager:viewviewAllows read-only access to Voyager CRDs, intended to be granted within a namespace using a RoleBinding.

These user facing roles supports ClusterRole Aggregation feature in Kubernetes 1.9 or later clusters.

Using kubectl

Since Voyager uses its own TPR/CRD, you need to use full resource kind to find it with kubectl.

# List all voyager ingress
$ kubectl get ingress.voyager.appscode.com --all-namespaces

# List voyager ingress for a namespace
$ kubectl get ingress.voyager.appscode.com -n <namespace>

# Get Ingress YAML
$ kubectl get ingress.voyager.appscode.com -n <namespace> <ingress-name> -o yaml

# Describe Ingress. Very useful to debug problems.
$ kubectl describe ingress.voyager.appscode.com -n <namespace> <ingress-name>

Detect Voyager version

To detect Voyager version, exec into the operator pod and run voyager version command.

$ POD_NAMESPACE=kube-system
$ POD_NAME=$(kubectl get pods -n $POD_NAMESPACE -l app=voyager -o jsonpath={.items[0].metadata.name})
$ kubectl exec -it $POD_NAME -n $POD_NAMESPACE voyager version

Version = 9.0.0
VersionStrategy = tag
Os = alpine
Arch = amd64
CommitHash = ab0b38d8f5d5b4b4508768a594a9d98f2c76abd8
GitBranch = release-4.0
GitTag = 9.0.0
CommitTimestamp = 2017-10-08T12:45:26