kube-fledged is a kubernetes add-on for creating and managing a cache of container images directly on the worker nodes of a kubernetes cluster. It allows a user to define a list of images and onto which worker nodes those images should be cached (i.e. pre-pulled). As a result, application pods start almost instantly, since the images need not be pulled from the registry.
kube-fledged provides CRUD APIs to manage the lifecycle of the image cache, and supports several configurable parameters to customize the functioning as per one's needs.
Table of contents
- Use cases
- Prerequisites
- Quick Install using YAML manifests
- Quick Install using Helm chart
- Quick Install using Helm operator
- Build and Deploy
- How to use
- How it works
- Configuration Flags for Kubefledged Controller
- Supported Container Runtimes
- Supported Platforms
- Built With
- Contributing
- License
Use cases
- Applications that require rapid start-up. For e.g. an application performing real-time data processing needs to scale rapidly due to a burst in data volume.
- Serverless Functions where functions need to react immediately to incoming events.
- IoT applications that run on Edge devices when the network connectivity between the edge and image registry is intermittent.
- If a cluster administrator or operator needs to roll-out upgrades to an application and wants to verify before-hand if the new images can be pulled successfully.
Prerequisites
- A functioning kubernetes cluster (v1.16 or above). It could be a simple development cluster like minikube or a large production cluster.
- Cluster-admin privileges to the kubernetes cluster.
- All master and worker nodes having the "kubernetes.io/hostname" label.
- git, make, go, docker engine (>= 19.03), openssl, kubectl, helm, gpg and gnu-sed installed on a local linux or mac machine. kubectl configured properly to access the cluster.
Quick Install using YAML manifests
These instructions install kube-fledged to a separate namespace called "kube-fledged", using YAML manifests and pre-built images in Docker Hub.
-
Clone the source code repository
$ mkdir -p $HOME/src/github.com/senthilrch $ git clone https://github.com/senthilrch/kube-fledged.git $HOME/src/github.com/senthilrch/kube-fledged $ cd $HOME/src/github.com/senthilrch/kube-fledged
-
Deploy kube-fledged to the cluster
$ make deploy-using-yaml
-
Verify if kube-fledged deployed successfully
$ kubectl get pods -n kube-fledged -l app=kubefledged $ kubectl get imagecaches -n kube-fledged (Output should be: 'No resources found')
Quick Install using Helm chart
-
Create the namespace where kube-fledged will be installed
$ export KUBEFLEDGED_NAMESPACE=kube-fledged $ kubectl create namespace ${KUBEFLEDGED_NAMESPACE}
-
Verify and install latest version of kube-fledged helm chart
$ helm repo add kubefledged-charts https://senthilrch.github.io/kubefledged-charts/ $ helm repo update $ gpg --keyserver keyserver.ubuntu.com --recv-keys 92D793FA3A6460ED (or) gpg --keyserver pgp.mit.edu --recv-keys 92D793FA3A6460ED $ gpg --export >~/.gnupg/pubring.gpg $ helm install --verify kube-fledged kubefledged-charts/kube-fledged -n ${KUBEFLEDGED_NAMESPACE} --wait
Quick Install using Helm operator
These instructions install kube-fledged to a separate namespace called "kube-fledged", using Helm operator and pre-built images in Docker Hub.
-
Clone the source code repository
$ mkdir -p $HOME/src/github.com/senthilrch $ git clone https://github.com/senthilrch/kube-fledged.git $HOME/src/github.com/senthilrch/kube-fledged $ cd $HOME/src/github.com/senthilrch/kube-fledged
-
Deploy the helm operator to a separate namespace called "kubefledged-operator" and kube-fledged to a separate namespace called "kube-fledged". If you need to deploy to a different namespace, export the variables OPERATOR_NAMESPACE and KUBEFLEDGED_NAMESPACE
$ make deploy-using-operator
-
Verify if kube-fledged deployed successfully
$ kubectl get pods -n kube-fledged -l app.kubernetes.io/name=kube-fledged $ kubectl get imagecaches -n kube-fledged (Output should be: 'No resources found')
Build and Deploy
These instructions will help you build kube-fledged from source and deploy it to a separate namespace called "kube-fledged". If you need to deploy it to a different namespace, edit the namespace field of the manifests in "kube-fledged/deploy" accordingly.
Build
-
Clone the source code repository
$ mkdir -p $HOME/src/github.com/senthilrch $ git clone https://github.com/senthilrch/kube-fledged.git $HOME/src/github.com/senthilrch/kube-fledged $ cd $HOME/src/github.com/senthilrch/kube-fledged
-
If you are behind a proxy, export the following ENV variables (UPPER case)
export HTTP_PROXY=http://proxy_ip_or_hostname:port export HTTPS_PROXY=https://proxy_ip_or_hostname:port
-
Build and push the docker images to registry (e.g. Docker hub)
$ export RELEASE_VERSION=<your_tag> $ export CONTROLLER_IMAGE_REPO=docker.io/<your_dockerhub_username>/kubefledged-controller $ export WEBHOOK_SERVER_IMAGE_REPO=docker.io/<your_dockerhub_username>/kubefledged-webhook-server $ export CRI_CLIENT_IMAGE_REPO=docker.io/<your_dockerhub_username>/kubefledged-cri-client $ export OPERATOR_IMAGE_REPO=docker.io/<your_dockerhub_username>/kubefledged-operator $ docker login -u <username> -p <password> $ export DOCKER_CLI_EXPERIMENTAL=enabled $ make install-buildx && make release-amd64
Deploy
Note:- You need to have 'cluster-admin' privileges to deploy
-
All manifests required for deploying kube-fledged are present in 'kube-fledged/deploy' directory.
-
Edit "kubefledged-deployment-controller.yaml".
Set "image" to "<your_docker_hub_username>/kubefledged-controller:<your_tag>"
image: <your_docker_hub_username>/kubefledged-controller:<your_tag>
-
If you pushed the image to a private repository, add 'imagePullSecrets' to the end of "kubefledged-deployment-controller.yaml". Refer to kubernetes documentation on Specifying ImagePullSecrets on a Pod. The secret <your_registry_key> should be created in "kube-fledged" namespace.
serviceAccountName: kubefledged imagePullSecrets: - name: <your_registry_key>
-
Edit "kubefledged-deployment-webhook-server.yaml".
Set "image" to "<your_docker_hub_username>/kubefledged-webhook-server:<your_tag>"
image: <your_docker_hub_username>/kubefledged-webhook-server:<your_tag>
-
Deploy kube-fledged to the cluster
$ make deploy-using-yaml
-
Verify if kube-fledged deployed successfully
$ kubectl get pods -n kube-fledged -l app=kubefledged $ kubectl logs -f <pod_name_obtained_from_above_command> -n kube-fledged $ kubectl get imagecaches -n kube-fledged (Output should be: 'No resources found')
How to use
kube-fledged provides APIs to perform CRUD operations on image cache. These APIs can be consumed via kubectl or curl
Create image cache
Refer to sample image cache manifest in "deploy/kubefledged-imagecache.yaml". Edit it as per your needs before creating image cache. If images are in private repositories requiring credentials to pull, add "imagePullSecrets" to the end.
imagePullSecrets:
- name: myregistrykey
Create the image cache using kubectl. Verify successful creation
$ kubectl create -f deploy/kubefledged-imagecache.yaml
$ kubectl get imagecaches -n kube-fledged
View the status of image cache
Use following command to view the status of image cache in "json" format.
$ kubectl get imagecaches imagecache1 -n kube-fledged -o json
Add/remove images in image cache
Use kubectl edit command to add/remove images in image cache. The edit command opens the manifest in an editor. Edit your changes, save and exit.
$ kubectl edit imagecaches imagecache1 -n kube-fledged
$ kubectl get imagecaches imagecache1 -n kube-fledged -o json
Refresh image cache
kube-fledged supports both automatic and on-demand refresh of image cache. Auto refresh is enabled using the flag --image-cache-refresh-frequency:
. To request for an on-demand refresh, run the following command:-
$ kubectl annotate imagecaches imagecache1 -n kube-fledged kubefledged.io/refresh-imagecache=
Delete image cache
Before you could delete the image cache, you need to purge the images in the cache using the following command. This will remove all cached images from the worker nodes.
$ kubectl annotate imagecaches imagecache1 -n kube-fledged kubefledged.io/purge-imagecache=
View the status of purging the image cache. If any failures, such images should be removed manually or you could decide to leave the images in the worker nodes.
$ kubectl get imagecaches imagecache1 -n kube-fledged -o json
Finally delete the image cache using following command.
$ kubectl delete imagecaches imagecache1 -n kube-fledged
Remove kube-fledged
Run the following command to remove kube-fledged from the cluster.
$ make remove-kubefledged (if you deployed using YAML manifests)
$ make remove-operator-and-kubefledged (if you deployed using Helm Operator)
How it works
Kubernetes allows developers to extend the kubernetes api via Custom Resources. kube-fledged defines a custom resource of kind “ImageCache” and implements a custom controller (named kubefledged-controller). kubefledged-controller does the heavy-lifting for managing image cache. Users can use kubectl commands for creation and deletion of ImageCache resources.
kubefledged-controller has a built-in image manager routine that is responsible for pulling and deleting images. Images are pulled or deleted using kubernetes jobs. If enabled, image cache is refreshed periodically by the refresh worker. kubefledged-controller updates the status of image pulls, refreshes and image deletions in the status field of ImageCache resource.
For more detailed description, go through kube-fledged's design proposal.
Configuration Flags for Kubefledged Controller
--image-pull-deadline-duration:
Maximum duration allowed for pulling an image. After this duration, image pull is considered to have failed. default "5m"
--image-cache-refresh-frequency:
The image cache is refreshed periodically to ensure the cache is up to date. Setting this flag to "0s" will disable refresh. default "15m"
--image-pull-policy:
Image pull policy for pulling images into and refreshing the cache. Possible values are 'IfNotPresent' and 'Always'. Default value is 'IfNotPresent'. Image with no or ":latest" tag are always pulled.
--stderrthreshold:
Log level. set the value of this flag to INFO
Supported Container Runtimes
- docker
- containerd
- cri-o
Supported Platforms
- linux/amd64
- linux/arm
- linux/arm64
Built With
- kubernetes/sample-controller - Building our own kubernetes-style controller using CRD.
- kubebuilder - SDK for building Kubernetes APIs using CRDs
- operator-sdk - SDK for building Kubernetes applications
- cri-tools - CLI and validation tools for Kubelet Container Runtime Interface (CRI).
- buildx - Docker CLI plugin for extended build capabilities with BuildKit
- Go Modules - Go Modules for Dependency Management
- Make - GNU Make
Contributing
Please read CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests.
License
This project is licensed under the Apache 2.0 License - see the LICENSE.md file for details