Adding Existing External Nodes

Updated at：2025-10-27

To add purchased instances to a CCE cluster as worker nodes or to rejoin removed worker nodes to a node group, you can add the nodes manually via the Cloud Container Engine (CCE) console. After addition, the nodes will be managed consistently at the node group level. Make sure to review the usage limits and precautions before proceeding.

Prerequisites

Successfully create CCE cluster.
The node group has been successfully created.

Usage restrictions

Only BCC, BBC, and EBC instances under the same account can be added, and they must be in the same region and VPC as the cluster.
Server instances that are already part of other CCE clusters or node groups cannot be added.
Only servers with the same instance specifications (CPU, memory, GPU card count) as the current node group can be added, and the billing type must match the node group.
Only servers of the same type as the current node group can be added.
Only node groups with auto scaling disabled support adding existing nodes. If auto scaling is enabled for the node group, it must be disabled before adding existing nodes. For details, please refer to Node Group Management - Configure Auto Scaling.
Ensure your cluster node quota is sufficient. For CCE cluster quota limits, please refer to CCE Cluster Quota Limits.

Operation steps

Add existing server instances outside the cluster

Sign in to Cloud Container Engine Console (CCE).
Select Cluster List in the left navigation bar.
On the Cluster List page, click on the desired cluster to open the cluster management page.
Select Node Management > Node Group from the left navigation bar in the cluster management page.
In the target node group's operation column, select More > Add Existing Nodes to enter the Node Scale-up page; or click the target node group, enter the Node List page, and click Add Existing Nodes to enter the Node Scale-up page.
Click the Finish button to finish adding existing nodes. You can view added node information on the Node Group - Node List page.

ConfigMap	Required/Optional	Configuration
Node source	Required	This example selects "Outside the Cluster."
Node type	Required	The server type configured for the node group is selected by default and cannot be changed.
Node selection	Required	Select the server instances to add to the node group. Multiple selections are supported. Note: The number of nodes that can be added is constrained by the cluster size.
Reinstall OS	Required	Reinstalling the OS will reinstall the operating system on the selected instance. During this process, the data on the system disk will be erased and cannot be recovered (data on cloud disks will not be affected). Note: When nodes are added without reinstalling OS, old data on the server instance may cause node addition failures or other issues. It is recommended to reinstall OS.
Select a node	Required	You can configure the selected nodes to either reinstall the OS or join the cluster directly using existing configurations. This example selects Reinstall OS. Reinstalling the OS will erase all data on the system disk and it cannot be recovered (cloud disk data remains unaffected).
Disk mount	Optional	You can choose whether to inherit node group configurations. By default, configurations are not inherited. Do not inherit node group configuration: Choose whether to store container and image data on data disks or local disks. Check: Ensure the device name and mount directory are provided. Please note the following considerations: The system formats all selected node devices to ext4 and mounts the formatted devices directly to the directory, resulting in the loss of any previous data. Back up important data in advance. The mount settings you specify will apply to all nodes added in the current batch. Make sure the device names you provide are accurate. If the device name you provide does not match the actual name of the selected node, the system will report an error and terminate the initialization process for mismatched nodes. If the specified mount directory does not exist, the system will automatically create one. If the selected node is not currently mounted to a disk, it will not be automatically mounted to a new disk. To query the device name, please refer to View Disk Information. Unchecked: No disk mount options are configured, enabling you to mount disks manually or via scripts. Inherit node group configuration: No need to manually configure disk mounts; the system will automatically apply the node group's disk mount configuration to the selected nodes. Please take into consideration the following points: The selected nodes must match the node group’s instance specifications, disk count, disk type, and disk size to inherit its configuration. If they do not match, the selected nodes retain their original disk configurations and do not inherit those of the node group. When selected nodes match the node group model, the system will format the disks as ext4 file systems. Formatted disks will be mounted in reverse order, starting from the last disk. Formatting will erase existing data, so ensure you back it up beforehand. For existing BBC-type nodes added to a node group, inheriting the node group’s disk mount configurations is currently not supported. If none of the chosen nodes align with the node group, the node group configuration cannot be applied.
OS configuration	Required	When reinstalling the OS, you can decide whether to inherit the node group configuration or not. By default, the node group configuration is not inherited. If the OS reinstallation is canceled, the current OS configuration of the selected instance will remain unchanged. Inherit node group configuration: No need to manually select the OS or drivers. The system will automatically reinstall selected nodes according to the node group's active model OS and driver versions. Note: When inheriting node group configurations, the system will clear custom drivers (GPU Driver/CUDA/cuDNN, etc.) or "User Data Injection" configuration from selected instances. Do not inherit node group configuration: It require selecting the OS type and image version. Note: For GPU nodes, decide whether to install GPU drivers, CUDA, and cuDNN based on your requirements. If you enable custom driver installation, the selected versions of GPU drivers, CUDA, and cuDNN will be installed automatically. Otherwise, the instance will retain its existing custom drivers (GPU Driver/CUDA/cuDNN, etc.) or "User Data Injection" configurations.
Administrator user name	Required	The username used to log in to the server instance, with "root" being the default for Linux OS.
Login method	Required	Supports "Key Pair" and "Custom Password" options.
Inherit node group configuration.	Required	By default, this option is enabled. You can add additional configuration items for the nodes, such as worker security groups, node memory sharing, kubelet data directory, container data directory, deployment execution scripts, custom kubelet parameters, cordons, resource labels, and whether the IAM role uses the current node group configuration. If enabled, the node group configuration will automatically apply to added nodes. If disabled, you can customize the configurations to initialize nodes.

Notes:

Security groups are categorized into regular and enterprise types, which cannot be bound simultaneously. For details, please refer to Security Group Overview.

If the security group types are the same, the node group security group configuration will be added to the existing security group. Otherwise, the server's original security group will be automatically unbound, and the node group's security group configuration will be used.

K8s labels, taints, and annotations default to the current configuration of the node group.

Inheriting node group configurations will update the Baidu Cloud Compute instance's IAM role binding to match the node group's current configuration. Empty node group IAM role configurations will not unbind existing roles.

Add orphaned nodes within the cluster

Sign in to Cloud Container Engine Console (CCE).
Select Cluster List in the left navigation bar.
On the Cluster List page, click on the desired cluster to open the cluster management page.
Select Node Management > Node Group from the left navigation bar in the cluster management page.
In the target node group's operation column, select More > Add Existing Nodes to enter the Node Scale-up page; or click the target node group, enter the Node List page, and click Add Existing Nodes to enter the Node Scale-up page.
Click the Finish button to finish adding existing nodes. You can view added node information on the Node Group - Node List page.

ConfigMap	Required/Optional	Configuration
Node source	Required	The current example opts for "Inside the Cluster".
Node type	Required	The server type configured for the node group is selected by default and cannot be changed.
Node selection	Required	Select orphaned nodes within the cluster (not associated with any node group) and add them to the node group, supporting multiple selections. Note: The number of nodes that can be added is constrained by the cluster size.

Notes

After moving into a node group, original node labels , taints, annotations, node memory sharing configurations, and IAM role configurations will be updated to the node group's current configuration. Empty node group IAM role configurations will not unbind existing roles. Other configurations remain unchanged.

When existing instances are added to a node group, the original custom drivers (GPU Driver/CUDA/cuDNN) or “user data injection” configurations will be retained.

NodeGroup Introduction

Custom NodeGroup Kubelet Configuration

CCE CCE