Kubernetes Installation

8 min

You can use Kubernetes to install Manufacturing Connect Edge.

Before you begin

Get access to a configured Kubernetes cluster.

Get access to PersistentVolumeClaim (PVC). Learn more about Persistent Volumes from the Kubernetes documentation.

Get administrative permissions to run the kubectl command-line tool.

Obtain access to a Manufacturing Connect Edge service account .json file, which you can get from [email protected].

Depending on the specific requirements of your environment, you may need to get access to a storage class. Learn more about Storage Classes from the Kubernetes documentation.

For full DeviceHub driver support, ensure your Linux host kernel is version 6 or higher

To use Kubernetes to install Manufacturing Connect Edge, you will need to do the following:

Step 1: Launch a Manufacturing Connect Edge Kubernetes

Step 2: Log in to Manufacturing Connect Edge

To jump to a specific section, refer to the Table of Contents on the right side and click the corresponding step.

Step 1: Launch a Manufacturing Connect Edge Kubernetes

To launch a Manufacturing Connect Edge Kubernetes:

1. Request and download the service account .json file by sending an email to [email protected]. 2. Open a Command Prompt Terminal (CTRL+R). 3. Execute the following Linux commands to setup credentials for a Kubernetes namespace.

kubectl create namespace imc-edge-prod

The command creates a namespace named imc-edge-prod.

# Change SA file from string to Base64:
$encodedPassword = [Convert]::ToBase64String([Text.Encoding]::UTF8.GetBytes((Get-Content -Raw ~\<file_name>.json)))
# Step to add secret:
kubectl create secret docker-registry imc-credential --namespace imc-edge-prod --docker-server=us-docker.pkg.dev --docker-username=_json_key --docker-password="$encodedPassword"

Important: If you encounter an error as shown in the screenshot below when using Windows PowerShell, make sure to follow these steps:

The namespace has been created

There are no spaces in the command

Replace the double quotes with single quotes

Refer to the command parameters:

kubectl create secret sets a Secret for use with a Docker registry.

--namespace imc-edge-prod sets up a Kubernetes namespace of imc-edge-prod.

docker-registry imc-credential sets up Secret name as imc-credential. This should not be changed.

--docker-server=us-docker.pkg.dev sets the Docker registry server as us-docker.pkg.dev. This should not be changed.

--docker-username=_json_key sets the Docker registry username to connect to the server above. This should not be changed.

--docker-password="$(cat ~/<file_name>.json)" retrieves the password within the service account file requested from step 1.

Note: imc-edge-prod should be changed to an isolated namespace to suit deployment best.

Example: if the .json file is named abc123.json, execute: kubectl create secret --namespace imc-edge-prod docker-registry imc-credential --docker-server=us-docker.pkg.dev --docker-username=_json_key --docker-password="$(cat ~/abc123.json)"

4. Copy or download the following Manufacturing Connect Edge deployment file.

mc-edge-deployment.yaml

# Create ServiceAccount
apiVersion: v1
kind: ServiceAccount
metadata:
  name: imc-edge-docker
  namespace: imc-edge-prod
  labels:
    app.kubernetes.io/name: imc-edge-docker
    app.kubernetes.io/instance: imc-edge-instance
---
# Create PVC
kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: imc-edge-data
  namespace: imc-edge-prod
  labels:
    app.kubernetes.io/name: imc-edge-docker
    app.kubernetes.io/instance: imc-edge-instance
spec:
  accessModes:
    - "ReadWriteOnce"
  resources:
    requests:
      storage: "10G"
---
# Create Service
apiVersion: v1
kind: Service
metadata:
  name: imc-edge-docker
  namespace: imc-edge-prod
  labels:
    app.kubernetes.io/name: imc-edge-docker
    app.kubernetes.io/instance: imc-edge-instance
spec:
  type: ClusterIP
  ports:
    - port: 443
      targetPort: https
      protocol: TCP
      name: https
  selector:
    app.kubernetes.io/name: imc-edge-docker
    app.kubernetes.io/instance: imc-edge-instance
---
# Create Deployment
apiVersion: apps/v1
kind: Deployment
metadata:
  name: imc-edge-docker
  namespace: imc-edge-prod
  labels:
    app.kubernetes.io/name: imc-edge-docker
    app.kubernetes.io/instance: imc-edge-instance
spec:
  replicas: 1
  selector:
    matchLabels:
      app.kubernetes.io/name: imc-edge-docker
      app.kubernetes.io/instance: imc-edge-instance
  template:
    metadata:
      labels:
        app.kubernetes.io/name: imc-edge-docker
        app.kubernetes.io/instance: imc-edge-instance
    spec:
      serviceAccountName: imc-edge-docker
      securityContext:
        {}
      imagePullSecrets:
        - name: imc-credential
      containers:
        - name: imc-edge-docker
          securityContext:
            capabilities:
              add:
              - NET_ADMIN
          image: "us-docker.pkg.dev/litmus-customer-facing/imce-build/mce:latest"
          imagePullPolicy: IfNotPresent
          ports:
            - name: https
              containerPort: 443
              protocol: TCP
          volumeMounts:
            - mountPath: /var
              mountPropagation: 
              name: imc-edge-data
              subPath: 
              subPathExpr: 
          resources:
            requests:
              cpu: 1
              memory: 1024Mi
            limits:
              cpu: 4
              memory: 4096Mi
      volumes:
        - name: imc-edge-data
          persistentVolumeClaim:
            claimName: imc-edge-data

5. Execute following command to deploy Manufacturing Connect Edge Kubernetes cluster using the deployment file from Step 4.

kubectl apply -f mc-edge-deployment.yaml --namespace imc-edge-prod

Refer to the command parameters:

-f mc-edge-deployment.yaml uses the file mc-edge-deployment.yaml for deployment. This should not be changed.

--namespace imc-edge-prod sets imc-edge-prod to be the deployment that Kubernetes will use.

Special and Additional Considerations:

Manufacturing Connect Edge can function in both single/multi Kubernetes node clusters, and can support various networking modes. Your Kubernetes admin has the responsibility to configure proper networking.

Adjust CPU/RAM/Storage for the Kubernetes cluster as needed. Storage should be set to at least 10GB. Scale additional computing resources when using high-computing power Manufacturing Connect Edge features such as PLC drivers, CC, and Flows.

Consult Kubernetes official documentation for best practices.

Ensure credentials are stored properly with Kubernetes management.

NET_ADMIN permissions are required.

Step 2: Log in to Manufacturing Connect Edge

See Access the Manufacturing Connect Edge Web UI for details on logging in.

See Access the Terminal User Interface for instructions on how to navigate and log in for the first time.

Note: To identify the port configured by the admin, use the following command: kubectl get svc -n imc-edge-prod