DocsICP FilingConsole
Documentation
Home Page Intelligent Computing LINGJUN Cluster Creation and Management Manage Lingjun Clusters and Nodes

Manage Lingjun Clusters and Nodes

Updated at:
Copy as MD
Product
My Favorites

A Lingjun cluster is a collection of high-performance Lingjun compute nodes equipped with Lingjun optimization components. Each Lingjun node corresponds to a GPU server and can be used to deploy heterogeneous computing services. This topic describes how to manage Lingjun clusters and Lingjun nodes, including viewing cluster info, viewing node details, and scaling out a Lingjun cluster.

Manage Lingjun Clusters

image

The status of a Lingjun cluster can be one of the following:

  • Initialization failed.: To view the details of the failed job, see O&M Task Center.

  • Initializing: Configuring the Lingjun network and initializing Lingjun compute nodes.

  • Running: You can scale out or scale in a cluster, or reinstall or restart a node only when the cluster is in the Running state.

    Important

    If the cluster scale-out, cluster scale-in, node reinstallation, and node restart jobs involve different Lingjun compute nodes, you can submit them in parallel.

View Cluster Info

  1. Log on to the Lingjun console.

  2. In the left-side navigation bar, choose Clusters & Nodes > Cluster Management.

  3. Click Details next to the cluster ID to go to the Cluster Details page.

    1. View basic cluster information, such as the cluster name, number of groups, and creation information.

    2. View cluster information across five tabs: Node Group, Monitoring Overview, Basic metrics, RDMA, and GPU.

Scale Out Cluster

Note

  • When you scale out a cluster, deploy the CPFS client on each new GPU node and add the node to the CPFS cluster.

  • When scaling out a cluster, you must tag the newly added nodes.

  1. Log on to the Lingjun console.

  2. In the left-side navigation bar, choose Clusters & Nodes > Cluster Management.

  3. Find the target cluster and click Actions in the Scale out column.

    1. In the Original Group Details: section, click Join Existing Node next to the corresponding node group name.

    2. In the dialog box that appears, enter the edge zone name prefix, logon password, and confirm password.

    3. Select the check box next to one or more unused edge zone instances or purchase new edge zones, then click Yes.

  4. In the Scale-out configuration area, click Confirm Submission.

  5. Return to the Cluster Management page. The cluster status shows Scaling out. Wait for the scale-out to complete.

Scale In a Cluster

Warning

  • Scaling in a cluster triggers node reinstallation for removed nodes and deletes all local data on those nodes. Therefore, back up node data before performing a scale-in operation.

  • During cluster scale-in, the nodes to be scaled in must be removed from the CPFS cluster.

  1. Log on to the Lingjun console.

  2. In the left-side navigation bar, choose Clusters & Nodes > Cluster Management.

  3. In the Actions column, click scale-in.

    1. In the Original Group Details: section, select the check box next to each node you want to remove, then click Batch Remove from Cluster.

    2. In the The following information displays the detailed configurations for scale-down: section, click Confirm Submission.

  4. On the Confirm scale-in page, enter DELETE in the text box and click Yes to scale in the cluster.

  5. Return to the Cluster Management page. The cluster status shows Scaling in. Wait for the scale-in to complete.

Delete Cluster

Important

  • Before deleting a cluster, you must scale in the cluster to remove all nodes.

  • The associated CPFS cluster is not deleted when you delete a cluster.

  1. Log on to the Lingjun console.

  2. In the left navigation bar, choose Clusters & Nodes > Cluster Management.

  3. Click the Cluster ID/Name you want to delete. On the Cluster Details page, click Delete in the upper-right corner.

  4. In the dialog box that appears, click Yes to delete the cluster.

Create a Cluster Group

You can create a cluster group for a Lingjun cluster in one of the following ways:

  • Create a cluster group when you create the cluster. For more information, see Cluster and group configuration.

  • After the cluster is created, create a group for an existing cluster.

    1. Log on to the Lingjun console.

    2. In the left-side navigation bar, choose Clusters & Nodes > Cluster Management.

    3. Click the corresponding Cluster ID/Name.

    4. Click the Node Group tab.

    5. Click Create Group. Enter the group name, default instance type, and other info for the node group.

    6. Optional. After you create a cluster group, you can edit its name or delete it.

Manage Lingjun nodes

Important

A Lingjun compute node can execute only one operation at a time, including cluster scale-out, cluster scale-in, node reinstallation, and node restart.

Purchase Node

  1. Log on to the Lingjun console.

  2. In the left-side navigation bar, choose > Node Management.

  3. On the Node Management page, click Purchase Node to go to the node purchase page.

  4. Follow the on-screen instructions to purchase a node.

View node details

  1. Log on to the Lingjun console.

  2. In the left-side navigation bar, choose Clusters & Nodes > Node Management to open the Node Management page.

  3. Click the All Nodes tab to view all nodes.

    • You can view the basic information about a node, such as the Node ID / Hostname, Image Name, Zone, etc.

    • Select conditions such as Image Name, Zone, or IP Address from the dropdown list, then enter search content in the text box to search for nodes.

  4. Click the Unused tab to view unused nodes. You can view the basic information about an unused node, such as Node Model and GPU.

Logon Node

  1. Log on to the Lingjun console.

  2. In the left-side navigation bar, choose Clusters & Nodes > Node Management Node Management

  3. Find the target node, and click Remote Logon in the Actions column.

    • The logon username is root.

    • The logon password is the cluster's logon password.

Reinstall Node

Important

  • Reinstalling a node will delete the node data. Exercise caution when performing this operation.

  • You can reinstall a node only when the Lingjun cluster is in the Running status.

  • When reinstalling a node, you must first remove the old node from the CPFS cluster and then add the new node information to the CPFS cluster.

Reinstall a node in the following situations:

  • Redeploy the business.

  • Change the operating system version.

  • Required for O&M scenarios.

Steps:

  1. Log on to the Lingjun console.

  2. In the left-side navigation bar, choose Clusters & Nodes > Node Management.

  3. On the Node Management page, click Reinstall next to the instance ID. In the dialog box that appears, select a runtime image version, modify the node name, enter and confirm the root password of the node, and then click Reinstall.

Restart Edge Zone

Important

  • Restarting an edge zone may affect business continuity.

  • You can restart an edge zone only when the Lingjun cluster status is Running.

You need to restart an edge zone in the following situations:

  • Deploy a new application or service.

  • Modify system configuration.

  • O&M scenario requirements.

Steps:

  1. Log on to the Lingjun console.

  2. In the left-side navigation bar, choose Clusters & Nodes > Node Management Node Management

  3. On the Node Management page, click Restart next to the instance ID.

Previous：Create Lingjun ClusterNext：Cluster management and O&M