Preparing to install the extension on a managed Kubernetes cluster (AKS) in Azure
The CloudPoint cloud-based extension can be deployed on a managed Kubernetes cluster in Azure for scaling the capacity of the CloudPoint host to service a large number of requests concurrently.
Overview
Your Azure managed Kubernetes cluster should already be deployed with appropriate network and configuration settings, and with specific roles. The cluster must be able to communicate with CloudPoint.
The required roles are:
Azure Kubernetes Service RBAC Writer, AcrPush, Azure Kubernetes Service Cluster User RoleFor supported Kubernetes versions, refer to the CloudPoint Hardware Compatibility List (HCL).
Use an existing Azure Container Registry or create a new one, and ensure that the managed Kubernetes cluster has access to pull images from the container registry
A dedicated nodepool for CloudPoint workloads needs to be created with manual scaling or 'Autoscaling' enabled in the Azure managed Kubernetes cluster. The autoscaling feature allows your nodepool to scale dynamically by provisioning and de-provisioning the nodes as required automatically.
CloudPoint extension images (
flexsnap-cloudpoint, flexsnap-listener, flexsnap-workflow, flexsnap-fluentd, flexsnap-datamover) need to be uploaded to the Azure container registry.
Prepare the host and the managed Kubernetes cluster in Azure
Choose the CloudPoint image supported on Ubuntu or RHEL system that meets the CloudPoint installation requirements and create a host.
See Creating an instance or preparing the host to install CloudPoint.
It is not recommended to scale the cluster up or down when a job is running. It might cause the job to fail. Set the cluster size beforehand.
Verify that the port 5671 is open on the main CloudPoint host.
See Verifying that specific ports are open on the instance or physical host.
The public IP of the virtual machine scale set via which the node pool is configured has to be allowed to communicate through port 22, on the workloads being protected.
Install a Docker or Podman container platform on the host and start the container service.
Prepare the CloudPoint host to access Kubernetes cluster within your Azure environment.
Install Azure CLI.
Install Kubernetes CLI
Login to the Azure environment to access the Kubernetes cluster by running this command on Azure CLI:
# az login --identity
# az account set --subscription <subscriptionID>
# az aks get-credentials --resource-group <resource_group_name> --name <cluster_name>
Ensure to create an Azure Container Registry or use the existing one if available, to which the CloudPoint images will be pushed (uploaded). See Azure documentation:
To run the kubectl and container registry commands from the host system, assign the following role permissions to your VM and cluster. You can assign a 'Contributor', 'Owner', or any custom role that grants full access to manage all resources.
Go to your Virtual Machine > click Identity on the left > under System assigned tab, turn the Status to 'ON' > click Azure role assignment > click Add role assignments > select Scope as 'Subscription' or 'Resource Group' > select Role and assign the following roles : Azure Kubernetes Service RBAC Writer, AcrPush, Azure Kubernetes Service Cluster User Role, and Save.
Go to your Kubernetes cluster > click Access Control (IAM) on the left > click Add role assignments > select Role as 'Contributor ' > Select Assign access to as 'Virtual Machines' > select your VM from the drop-down and Save.
Create a storage account in the same subscription and region your Kubernetes cluster is in, and create a file share into it. (Follow the default settings by Azure.)
Create a namespace for CloudPoint from the command line on host system:
# kubectl create namespace cloudpoint-system
Then create a new or use an existing managed Kubernetes cluster in Azure, and add a new node pool dedicated for CloudPoint use. Configure Autoscaling as per your requirement.
Ensure that Azure plug-in is configured.
