Add Node
The upper limit of Node for a single cluster is 5,000, and up to 50 Nodes can be added at one time on the console page. On the cluster details page and the node list page, click “Add Node”.
Alternatively, go to the Node Pool page, select a node pool, and click the “Add Node” button to add nodes to the node pool.
| Configuration Item | Description |
|---|---|
| Subnet | Set the subnet where the initial node and Pod are located. Nodes in the cluster can be in different subnets under the same VPC. |
| Type | Node type, supporting UHost cloud server (including GPU cloud server),. |
| Availability Zone | Master/Node’s availability zone. In regions with multiple availability zones, you can choose multi-availability zone UK8S cluster. It is recommended to distribute the Master nodes in multiple availability zones when creating the cluster. |
| Node Image | Set the UHost image of the cluster node. You can choose a custom image, but it must be based on the UK8S standard image production. For details, please refer to Create Custom Image If you want to use a GPU node, please refer to Image Description in GPU Node. CPU machine image can be chosen from: Centos 7.6, Ubuntu 20.04, Anolis 8.6 images. |
| Node Specifications | Including machine type, CPU platform, CPU, memory, system disk type, data disk type, data disk size, etc. The configuration of different types of models is detailed in the corresponding cloud product documentation |
| Hardware Isolation Group | Master nodes are by default in the same hardware isolation group, and the hardware isolation group can strictly ensure that each cloud server within the group falls on different physical machines. Each isolation group can add up to 7 cloud servers in a single availability zone. For details, see Hardware Isolation Group |
| Maximum Pod Count | The maximum number of Pods that a single Node can support |
| Label | Node label. See Kubernetes official documentation: Labels and Selectors |
| Taint | Node taint. See Kubernetes official documentation: Taints and Tolerations |
| Disable Nodes | Enabling the “Disable Node (cordon node)” option will add the node to the cluster as an unschedulable node, which can be enabled as needed. |
| Custom Data | Refers to the configuration script that the system automatically runs when the host starts for the first time or each time. This script can be passed into the metadata server by console API and other means, and obtained by the cloud-init program inside the host. The script follows the standard CloudInit syntax. This script will block the UK8S install script, i.e., the install of K8S related components such as Kublet, Scheduler, etc. will only start after this script is executed. |
| Initial Script | This script executes only once after UK8S starts, and runs after K8S-related components are successfully installed. Following standard shell syntax, execution results are stored in the /var/log/message/ directory. |
Node Operations
After adding nodes, you can view the node list on the node list page of the cluster page. You can also perform operations such as “Disable”, “Delete”, “Details”, “Bind EIP”, and “Modify External Firewall” on the nodes.
Disabling
Disabling a node executes the kubectl cordon node command, making the node Unschedulable without evicting existing Pods on the node.
Deletion
When deleting a node, you can check the option to simultaneously delete the cloud host resources and data disks on the cloud host. If not checked, the node will only be removed from the cluster.
Recommendation: Evict services from the node before deleting it to ensure service availability.
Node Labels
You can add, modify, or delete labels for Worker nodes via the Modify Labels button on the node list page. Node labels are key-value pairs in Kubernetes used to identify node characteristics, applicable for Pod scheduling policies.
Label Format Requirements
- Represented as key-value pairs (key=value).
- Each operation supports modifying up to** 20** label pairs.
- Label key restrictions:
- Avoid using kubernetes.io or k8s.io as prefixes, as these are reserved for Kubernetes system labels.
- Using system-reserved prefixes may cause scheduling exceptions or conflicts with system labels.
- Do not modify labels added by the system by default, as modifications may lead to system service issues.
- Both keys and values must comply with Kubernetes standard naming conventions.
Operation Steps
- In the node list, select More Operations for the target node.
- Click Modify Labels.
- Add or modify labels in the pop-up dialog box.
- Click OK to complete the modification.
Node Taints
You can add, modify, or delete taints for Worker nodes via the Modify Taints button on the node list page. Node taints control which Pods can be scheduled onto a node and must be used in conjunction with Pod tolerations .
Taint Format Requirements
- Represented as key=value:effect.
- Each operation supports updating up to 10 taint pairs.
- effect must be one of the three types:
- NoSchedule: Prevents new Pods from being scheduled to the node (existing Pods remain unaffected).
- PreferNoSchedule: Discourages, but does not strictly prevent, new Pods from being scheduled to the node.
- NoExecute: Prevents new Pods from being scheduled and evicts existing Pods that do not tolerate the taint.
Operation Steps
- In the node list, select More Operations for the target node.
- Click Modify Taints.
- Add or modify taints in the pop-up dialog box.
- Click OK to complete the modification.