serving-operator

module
v0.12.0 Latest Latest
Warning

This package is not in the latest version of its module.

Go to latest
Published: Jan 23, 2020 License: Apache-2.0

README

Knative Serving Operator

Go Report Card Releases LICENSE Slack Status

Knative Serving Operator is a project aiming to deploy and manage Knative Serving in an automated way.

The following steps will install Knative Serving and configure it appropriately for your cluster in the knative-serving namespace. Please make sure the prerequisites are installed first.

  1. Install the operator
  • Installing from source code:

To install from source code, run the command:

ko apply -f config/
  • Installing a released version:

To install a released version of the operator go and download the latest serving-operator.yaml file from here and apply it (kubectl apply -f serving-operator.yaml), or directly run:

kubectl apply -f https://github.com/knative/serving-operator/releases/download/v0.10.0/serving-operator.yaml
  1. Install the KnativeServing custom resource
cat <<-EOF | kubectl apply -f -
apiVersion: v1
kind: Namespace
metadata:
 name: knative-serving
---
apiVersion: operator.knative.dev/v1alpha1
kind: KnativeServing
metadata:
  name: knative-serving
  namespace: knative-serving
spec:
  config:
    defaults:
      revision-timeout-seconds: "300"  # 5 minutes
    autoscaler:
      stable-window: "60s"
    deployment:
      registriesSkippingTagResolving: "ko.local,dev.local"
    logging:
      loglevel.controller: "debug"
EOF

Please refer to Building the Operator Image to build your own image.

Prerequisites

Istio

On OpenShift, Istio will get installed automatically if not already present by using the Maistra Operator.

For other platforms, see the docs

Operator SDK

This operator was originally created using the operator-sdk. It's not strictly required but does provide some handy tooling.

The KnativeServing Custom Resource

The installation of Knative Serving is triggered by the creation of a KnativeServing custom resource (CR) as defined by this CRD. The operator will deploy Knative Serving in the same namespace containing the KnativeServing CR, and this CR will trigger the installation, reconfiguration, or removal of the knative serving resources.

The optional spec.config field can be used to set the corresponding entries in the Knative Serving ConfigMaps. Conditions for a successful install and available deployments will be updated in the status field, as well as which version of Knative Serving the operator installed.

The following are all equivalent:

kubectl get knativeservings.operator.knative.dev -oyaml
kubectl get knativeserving -oyaml
kubectl get ks -oyaml

To uninstall Knative Serving, simply delete the KnativeServing resource.

kubectl delete ks --all

Development

It can be convenient to run the operator outside of the cluster to test changes. The following command will build the operator and use your current "kube config" to connect to the cluster:

./hack/run-local.sh

Pass --help for further details on the various subcommands

Building the Operator Image

To build the operator with ko, configure your an environment variable KO_DOCKER_REPO as the docker repository to which developer images should be pushed (e.g. gcr.io/[gcloud-project], docker.io/[username], quay.io/[repo-name], etc).

Install ko with the following command, if it is not available on your machine:

go get -u github.com/google/ko/cmd/ko

Then, build the operator image:

ko publish knative.dev/serving-operator/cmd/manager -t $VERSION

You need to access the image by the name KO_DOCKER_REPO/manager-[md5]:$VERSION, which you are able to find in the output of the above ko publish command.

The image should match what's in config/operator.yaml and the $VERSION should match version.go and correspond to the contents of config/.

Operator Lifecycle Manager and OperatorHub

Knative Serving operator has the metadata in Operator Lifecycle Manager (OLM) checked in at deploy/olm-catalog. Files in there are for reference purposes and also for testing and tooling.

In order to install the operator CatalogSource to a cluster with OLM, run these commands:

OLM_NS=$(kubectl get deploy --all-namespaces | grep olm-operator | awk '{print $1}')
./hack/generate-olm-catalog-source.sh | kubectl apply -n $OLM_NS -f -

Then install the operator by creating a subscription:

OLM_NS=$(kubectl get operatorgroups --all-namespaces | grep olm-operators | awk '{print $1}')
OPERATOR_NS=$(kubectl get operatorgroups --all-namespaces | grep global-operators | awk '{print $1}')

cat <<-EOF | kubectl apply -f -
apiVersion: operators.coreos.com/v1alpha1
kind: Subscription
metadata:
  name: knative-serving-operator-sub
  generateName: knative-serving-operator-
  namespace: $OPERATOR_NS
spec:
  source: knative-serving-operator
  sourceNamespace: $OLM_NS
  name: knative-serving-operator
  channel: alpha
EOF

Upgrade

Please refer to the upgrade guide for a safe upgrade process.

Directories

Path Synopsis
cmd
pkg
client/clientset/versioned
This package has the automatically generated clientset.
This package has the automatically generated clientset.
client/clientset/versioned/fake
This package has the automatically generated fake clientset.
This package has the automatically generated fake clientset.
client/clientset/versioned/scheme
This package contains the scheme of the automatically generated clientset.
This package contains the scheme of the automatically generated clientset.
client/clientset/versioned/typed/serving/v1alpha1
This package has the automatically generated typed clients.
This package has the automatically generated typed clients.
client/clientset/versioned/typed/serving/v1alpha1/fake
Package fake has the automatically generated clients.
Package fake has the automatically generated clients.
e2e

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL