All Products
Search

This Product

    Platform For AI:Create an ACK Lingjun managed cluster

    Document Center

    Platform For AI:Create an ACK Lingjun managed cluster

    Last Updated:Mar 11, 2026

    Create an ACK Lingjun managed cluster with a fully managed control plane. The cluster supports Lingjun compute nodes as worker nodes.

    Prerequisites

    • Purchase required Intelligent Computing LINGJUN products, such as compute nodes and Lingjun connection instances. For more information, see Purchase products.

    • Purchase and configure required cloud products, such as Cloud Enterprise Network (CEN), Application Real-Time Monitoring Service (ARMS), and Virtual Private Cloud (VPC). For more information, see Purchase and configure other cloud products.

    • Complete identity verification. Maintain a cash or credit balance of at least CNY 100 to create ACK clusters.

    Background information

    ACK Lingjun managed clusters provide a standard Kubernetes service with a fully managed, high-availability control plane. These clusters support efficient management of heterogeneous resources and scheduling of heterogeneous tasks. They serve as the cloud-native foundation for PAI and provide enhanced cloud-native capabilities for high-performance computing (HPC) scenarios such as AI. For more information about ACK Lingjun managed clusters, see What is an ACK Lingjun managed cluster.

    Create a cluster configuration

    1. Log on to the Intelligent Computing LINGJUN console.

    2. In the navigation pane on the left, choose Resources & Nodes > Cluster Management to go to the Cluster Management page.

    3. Click Create Cluster.

    4. Click the PAI LINGJUN Cluster (including PAI, ACK, and CPFS) card.

      Configure the cluster and groups, network, basic parameters for software instances, and mappings between software instances and groups.

    Note

    ACK Lingjun managed clusters are billed separately. For more information, see Billing rules.

    Configure cluster and node groups

    Plan clusters and organize compute nodes into node groups to improve resource utilization.

    image

    1. Configure cluster information.

      Enter cluster name, root password for cluster nodes, and resource group. For more information, see Create a resource group.

    2. Click Create Group to configure node group information.

      1. Configure group name and node information: node model and runtime image.

      2. Click Select Node Instances to choose nodes to add to this group.

    3. Click Save and Next: Network Configuration.

    Configure network

    By default, clusters run in isolated network environments. To connect a cluster to public cloud, use a Lingjun connection instance and CEN instance, and specify a VPC to monitor network connectivity.

    image

    Network configuration involves these networks:

    • Cluster CIDR block: Private CIDR block to assign IP addresses to compute nodes.

    • Monitoring network: VPC to monitor network connectivity.

    Ensure CIDR blocks do not overlap.

    Note

    After configuring cluster network, verify CEN network configuration. For more information, see Purchase and configure CEN.

    1. Configure the cluster CIDR block.

      • Enter a valid private CIDR block to assign IP addresses to compute nodes.

      • The cluster subnet is a subnet of the cluster CIDR block. For more information, see Manage Lingjun CIDR blocks.

      Note

      • Cluster CIDR block cannot overlap with other connected networks: VPCs or on-premises data centers.

      • Available IP addresses in the cluster CIDR block determine maximum nodes. Reserve a sufficiently large CIDR block (mask length less than 22 recommended) for future scaling.

    2. (Optional) Configure cluster subnet bond allocation policy. This policy configures allocation of bond interfaces on physical network cards. Configure bond interfaces using a bond policy, model policy, or node policy.

      Add a bond policy

      Node models have different numbers of bond interfaces. The cluster bond interface count is determined by the model with the most bond interfaces. Bond interfaces are named bondx, where x uses zero-based numbering.

      For example, if a cluster has two node models with 3 and 4 bond interfaces, the cluster has 4 bond interfaces: bond0 through bond3. The 3-interface model uses policies for bond0 through bond2.

      Note

      Configure only one bond allocation policy per cluster.

      Steps

      1. Configure the bond policy for the cluster.

      2. (Optional) Configure the default bond. This policy applies to bond interfaces without specific policies. Select Apply to all to apply the default policy to all bond interfaces.

      Add a model policy

      Specify a model allocation policy for each node model. Create up to the number of groups in the cluster.

      Procedure

      1. Select the node model from the drop-down list.

      2. Configure the bond policy. The policy applies to all node instances of the selected model.

      Add a node policy

      Configure a node allocation policy for each node instance. Different bond ports of the same node can connect to different Lingjun CIDR blocks or subnets.

      Procedure

      1. Select the node instance from the drop-down list.

      2. Configure the bond policy. The policy applies to the selected node instance.

    3. Configure a Lingjun connection instance.

      1. Click Authorize to grant required permissions to the Lingjun connection instance.

        The Lingjun connection instance connects to CEN and accesses cloud products. Grant Lingjun permissions to access these products. For more information, see Appendix: Service-linked role for Lingjun connection instances.

      2. From the drop-down list, select the Lingjun connection instance ID for cluster connection to cloud.

      3. From the drop-down list, select the Cloud Enterprise Network instance to connect through the Lingjun connection instance.

        Important

        Create a transit router in CEN. The transit router region must match the Lingjun node region. For more information, see Transit router instances.

    4. Configure monitoring network information.

      1. Create a new VPC or connect an existing VPC to the transit router in the selected CEN instance. For more information, see Enable and configure CEN. Ensure the vSwitch in the VPC has at least one available IP address. Lingjun uses this vSwitch to monitor Lingjun connection instance connectivity.

        Important

        • Select a VPC from the drop-down list only after connecting the VPC to the selected transit router.

        • The cluster CIDR block and monitoring VPC CIDR block cannot overlap. The monitoring VPC CIDR block also cannot overlap with other connected networks: VPCs or on-premises data centers.

      2. Click the image icon next to the VPC and vSwitch drop-down lists. Select the VPC and vSwitch.

    5. Click Save and Next: Basic Software Instance Parameters.

    Configure basic parameters for software instances

    1. Configure the basic parameters.

      ACK tab

      Configure parameters for the ACK Lingjun managed cluster. For more information about cluster parameters, see Create an ACK managed cluster.

      Important

      The Service CIDR block, LINGJUN Cluster CIDR block, Internet CIDR block, and VPC CIDR block cannot overlap.

      CPFS tab

      Configure parameters for CPFS.

      Note

      After the CPFS file system is created, view instance information in the CPFS console.

      PAI tab

      Configure parameters for PAI.

      Note

      For more information about configuring ApsaraDB RDS, cloud storage, ACR image repositories, and OAuth authentication, see Activate and configure other cloud products.

    2. Click Save and Go to Next Step: Mappings Between Software Instances and Groups.

    Configure mappings between software instances and groups

    ACK Lingjun managed clusters provide Lingjun node pools for Lingjun compute nodes. This allows you to manage Lingjun nodes in groups more easily. For example, configure nodes, manage nodes in batches, specify scheduling, and configure GPUs. For more information about node pools, see Overview of Lingjun node pools.

    1. Click Create Node Pool to create an ACK node pool.

    2. Configure Node Pool Name, Maximum Number of Nodes, and other parameters.

    3. Click Select Associated Group. In the dialog box that appears, select the cluster groups to associate, and then click OK.

    4. Click Save and Go to Next Step: Confirm Configuration.

    Confirm configuration

    1. On the Confirm Configuration page, review the basic cluster information, network configuration, mappings between software instances and groups, and software instance parameters. Click Submit Configuration to start creating the cluster.

    2. In the Dependency Check section, click Complete Authorization to grant permissions to Container Service.