How to Add a Node Group to a Hosted Kubernetes Cluster?

Overview

Hosted Kubernetes allows you to add worker nodes to an existing cluster by creating a Node Group. A node group is a pool of worker nodes with a common configuration, including compute flavor, storage, security groups, key pair, autoscaling settings, labels, and taints.

Prerequisites

Before adding a node group:

Ensure that a Hosted Kubernetes cluster is available.
Ensure that you have the required quota available for the worker nodes, including the required compute flavor, number of worker nodes, and storage capacity
Ensure that you have the required Security Group and Key Pair, or permission to create them.
Determine the required compute flavor, number of worker nodes, and storage capacity.
If autoscaling is required, determine the minimum and maximum number of worker nodes.

Add a Node Group

Log in to the AceCloud portal.
Navigate to Kubernetes > Hosted > Overview.

Click the name of the cluster to which you want to add worker nodes.

Go to the Compute tab.

Under Node Groups, click Add Node Group.

Configure Node Group Details

Under Node Group Details, enter the following:

Node Group Name: Enter a unique name for the node group. This name is used as the hostname prefix for the worker nodes in the pool.

Configure Compute and Storage

Under Specification, configure the worker node resources:

Select the required Flavor Type.
Select the required Flavor.
Enter the Quantity of worker nodes to specify the number of worker nodes to create.
Enter the required Volume (GB) for worker node storage.

Note: GPU-based flavors are subject to availability. If you require a GPU flavor that is not currently available, create a Support Ticket through the AceCloud portal to request GPU resources.

Configure Access and Metadata

Under Access & Metadata, configure the following options:

Security Groups

Select the required Security Group for the worker nodes.

If the required security group is not available, click Create to create one.

Key Pair

Select the Key Pair that will be used for SSH access to the worker nodes.

If required, click Create to create a new key pair.

Configure Autoscaling

Enable Enable Autoscaling if you want the number of worker nodes in the node group to automatically scale based on the configured limits.

When autoscaling is enabled:

Minimum worker nodes: Specify the minimum number of worker nodes that the node group can scale down to.
Maximum worker nodes: Specify the maximum number of worker nodes that the node group can scale up to.

If autoscaling is disabled, the node group maintains the configured quantity of worker nodes.

Configure Worker Node Labels

Use Worker Node Labels to assign key-value labels to the worker nodes.

Enter a label Key.
Enter the corresponding Value.
Click + Add Label to add additional labels.

Labels can be used to identify and organize worker nodes and can be used by Kubernetes workloads for node selection.

Configure Taints

Use Taints to control which workloads can be scheduled on the worker nodes.

Enter the Key.
Enter the Value.
Select the required Effect.
Click + Add Taint to add additional taints.

Review the Node Group Configuration

The Node Group Overview panel displays a summary of the configuration, including:

Node Group Name
Flavor
Quantity of worker nodes
Volume
Key Pair
Security Groups
Worker Node Labels
Taints
Autoscaling status
Minimum worker nodes
Maximum worker nodes

Review the configuration before creating the node group.

Create the Node Group

Review the configuration in the Node Group Overview panel.
Click Create Node Group.

The node group is created and added to the selected Hosted Kubernetes cluster.

After successful creation, the node group and its worker nodes are available under the cluster's Compute tab.

Note: The available flavors, security groups, key pairs, and other options may vary depending on the project, region, and available resources.

Node States

After a node group is created or modified, worker nodes progress through different states during provisioning and removal. The State column in the node group displays the current status of each worker node.

State

Description

Pending

The node request has been accepted, and the node configuration is being prepared.

Provisioning

The required port, volume, and instance are being created.

Provisioned

The instance is booting and preparing to join the Kubernetes cluster.

Running

The worker node has successfully joined the Kubernetes cluster and is available for workloads.

Deleting

The port, volume, and instance associated with the worker node are being removed.

Note: The node state is updated as the provisioning or deletion process progresses. A newly added worker node may take some time to reach the Running state.

Manage Node Group Capacity

After creating a node group, you can manage its worker node capacity directly from the Compute tab. The node group provides the following actions:

Action

Description

Show node group details

Opens the node group details and displays its configuration, including autoscaling status, Kubernetes version, key pair, security groups, worker node labels, and taints.

Scale in

Removes one worker node from the node group, reducing the number of worker nodes in the pool by one.

Scale out

Adds one worker node to the node group, increasing the number of worker nodes in the pool by one.

Scale In

To remove a worker node:

In the required node group, click the – (Scale in) icon.
A confirmation dialog is displayed showing the node group scale in.
Review the details and click Confirm to proceed.

Note: Scaling in reduces the number of worker nodes in the node group by one. Ensure that removing a worker node does not affect your workload capacity.