Local Installation Test
In a production environment Shifu can be deployed directly in a K8s cluster.
Production deployment of Shifu is very easy, but for testing, we need to do some debugging work locally on the access device. This is where we need to test install Shifu locally and you need to create clusters locally, some tools like kind can help us to create such clusters on our own computers.
This article serves the same purpose as Quick Start, except that it breaks down the contents of the Shifu installer and creates a cluster locally using the command line step by step. This approach helps you understand the components used in the operation of Shifu.
Installing Docker Desktop
Docker Desktop can install Docker on desktop operating systems (Windows / macOS / Ubuntu for Desktop, etc.) and provides a visual interface for management.
Shifu uses Docker to turn every actual physical device (edgeDevice) into a digital twin (deviceShifu), the role of Docker is mainly virtualization and isolation.
Please check the Docker official website to install Docker Desktop on your own computer.
Make sure Docker Desktop is installed and started
Use the following command to make sure that Docker Desktop is installed and started, and the following output will indicate success.
$ sudo docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
If the output is Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?, then Docker Desktop is not started; if the output is command not found, then Docker Desktop is not installed.
Install kubectl
kubectl is a command line tool for Kubernetes that helps you execute commands in a Kubernetes cluster. You can use kubectl to deploy applications, view and manage cluster resources, and view logs.
Please check the K8s website for the content to install kubectl.
Confirm that kubectl is installed.
$ kubectl version --client --output=yaml
clientVersion:
buildDate: "2022-08-23T17:44:59Z"
compiler: gc
......
Install kind
kind allows us to create a Kubernetes cluster locally for testing.
If Go is already installed, you can install it with the following command.
$ go install sigs.k8s.io/kind@v0.14.0
If Go is not installed, you can check the official documentation of kind to choose the installation method.
Confirm that kind is installed.
$ kind --version
kind version 0.14.0
Create a Cluster
We use kind to create the cluster.
$ sudo docker pull kindest/node:v1.24.0
$ sudo kind create cluster --image="kindest/node:v1.24.0"
Creating cluster "kind" ...
✓ Ensuring node image (kindest/node:v1.24.0) 🖼
✓ Preparing nodes 📦
✓ Writing configuration 📜
✓ Starting control-plane 🕹️
✓ Installing CNI 🔌
✓ Installing StorageClass 💾
Set kubectl context to "kind-kind"
The K8s cluster used for local testing will have K8s-related management content, Shifu controllers, etc. installed; the digital twin of the device will be created in it when the device is plugged in later.
Note: Make sure the cluster is created
$ sudo kubectl cluster-info --context kind-kind
Kubernetes control plane is running at https://127.0.0.1:52138
CoreDNS is running at https://127.0.0.1:52138/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy
$ sudo kind get clusters
kind
Note: Re-Create
In case of problems, you can delete the cluster and recreate it.
$ sudo kind delete cluster
Deleting cluster "kind" ...
$ sudo kind create cluster --image="kindest/node:v1.24.0"
Install Shifu
The installation of Shifu is very easy, pkg/k8s/crd/install/shifu_install.yml is the installation script, and can be installed with one click:
# clone Shifu repository
git clone https://github.com/Edgenesis/shifu.git
cd shifu
# Install Shifu in the cluster
sudo kubectl apply -f pkg/k8s/crd/install/shifu_install.yml
To learn more about user metrics we collect and how to disable it, please check User Metrics in Shifu.
Note: Download the Image in Advance
The mirrors used in k8s/crd/install/shifu_install.yml are quay.io/brancz/kube-rbac-proxy:v0.13.1 and edgehub/shifu-controller:nightly, so if you have problems downloading them, you can go ahead and Download the image to the local machine and import it into the cluster:
sudo docker pull quay.io/brancz/kube-rbac-proxy:v0.13.1
sudo kind load docker-image quay.io/brancz/kube-rbac-proxy:v0.13.1
sudo docker pull edgehub/shifu-controller:nightly
sudo kind load docker-image edgehub/shifu-controller:nightly
This method will take up local storage. You can use the command sudo docker rmi <image_id> to delete the local image when you are done using it.
Next Step
Congratulations! You have opened a cluster on your machine and installed Shifu in that cluster. Next you can try access devices in the cluster now!