DocsICP FilingConsole
官方文档
首页

Management node

更新时间：
复制 MD 格式
产品详情
我的收藏

You can manage these compute nodes by connecting remotely, sending commands, rebooting them, or releasing them.

Prerequisites

  • The cluster is in the Running state.

  • To create nodes, you must meet the following requirements:

Note

The Nodes page displays only compute nodes. It does not display management nodes or logon nodes.

Create nodes

  1. Go to the Cluster Details page.

    1. Log on to the E-HPC console.

    2. In the left part of the top navigation bar, select a region.

    3. In the left-side navigation pane, click Cluster.

    4. On the Cluster List page, find the cluster that you want to manage and click the Cluster Name.

  2. In the left-side navigation pane, choose Nodes > Node.

  3. Click Add Node. On the Add Node page, configure the following parameters.

    • Basic settings

      Parameter

      Description

      Destination Queue

      Select an existing queue in the cluster.

      Nodes

      The number of nodes to add.

    • Node configurations

      Parameter

      Description

      Select Node Type

      Only New Node is supported.

      Inter-node interconnection

      Select the network communication mode for the nodes.

      • VPC: Nodes communicate with each other over a VPC network.

      • eRDMA: Enables communication over an eRDMA network for nodes that use an instance type supporting Elastic RDMA Interface (ERI).

        Note

        Only some instance types support ERI. For more information, see eRDMA overview and Enable eRDMA on an enterprise-level instance.

      Use Preset Node Pool

      Select a reserved node pool. The system will automatically use unassigned nodes from the pool to create compute nodes, assigning their IP addresses and hostnames.

      Note

      Using a reserved node pool for scale-out enables you to quickly reuse pre-allocated resources. For more information, see Use reserved node pools in clusters.

      vSwitch

      Select the vSwitch for the nodes. The system allocates IP addresses to the nodes from the vSwitch's available CIDR block.

      Instance type Group

      Click Add an instance specification to select an instance type for the nodes.

      If auto scaling is disabled, you can add only one instance type. If auto scaling is enabled, you can add multiple instance types.

      Hostname Prefix

      The hostname prefix for the nodes.

      Hostname Suffix

      The hostname suffix for the nodes.

      Instance RAM role

      Assigns a RAM role to the nodes, granting them access to other Alibaba Cloud services.

      We recommend that you select the default AliyunECSInstanceForEHPCRole role created by the system.

  4. Select I understand that the deletion protection feature is enabled by default for added nodes and is not affected by the queue scaling configuration, and then click Confirm Add.

    Note

    New compute nodes are automatically initialized with the cluster software. Existing compute nodes are not affected.

    You can view the status of the added nodes in the node list. When a node enters the Running state, the scale-out operation is complete.

Reboot nodes

If a compute node has an issue, you can reboot it using one of two available methods:

  • Normal Reboot: Sends a reboot command to the node. The operating system terminates processes and performs a system reboot.

  • Force Reboot: This is equivalent to a hard power cycle and may cause data loss. Use this method only if a normal reboot fails.

Important

Rebooting a node stops the jobs running on it. Make sure that no jobs are running on the node before you reboot it.

  1. Go to the Cluster Details page.

    1. Log on to the E-HPC console.

    2. In the left part of the top navigation bar, select a region.

    3. In the left-side navigation pane, click Cluster.

    4. On the Cluster List page, find the cluster that you want to manage and click the Cluster Name.

  2. In the left-side navigation pane, choose Nodes > Node.

  3. In the node list, select one or more target nodes.

  4. Below the list, click Reboot.

  5. In the dialog box that appears, select a reboot method, and then click OK.

Connect to a node

Note

A Workbench remote session lasts 6 hours by default. After 6 hours of inactivity, the session automatically closes, and you must reconnect.

  1. Go to the Cluster Details page.

    1. Log on to the E-HPC console.

    2. In the left part of the top navigation bar, select a region.

    3. In the left-side navigation pane, click Cluster.

    4. On the Cluster List page, find the cluster that you want to manage and click the Cluster Name.

  2. In the left-side navigation pane, choose Nodes > Node.

  3. On the Nodes page, find the target node and click Connect in the Operation column.

  4. In the Connect dialog box, click Log On next to Connect with Workbench.

  5. In the Instance login dialog box that appears, enter your logon information. For parameter details, see Connect to an instance by using Workbench.

Send commands

To quickly perform maintenance tasks, such as installing software or running operations and maintenance scripts, use the Send Command feature.

  1. Go to the Cluster Details page.

    1. Log on to the E-HPC console.

    2. In the left part of the top navigation bar, select a region.

    3. In the left-side navigation pane, click Cluster.

    4. On the Cluster List page, find the cluster that you want to manage and click the Cluster Name.

  2. In the left-side navigation pane, choose Nodes > Node.

  3. In the node list, select one or more target nodes.

  4. Below the list, click Send Command.

  5. In the dialog box that appears, set the command parameters and enter the command content.

    For parameter details, see Send remote commands.

  6. Click Run.

Deletion protection

Deletion protection is a security feature that prevents nodes from being accidentally deleted. When enabled, this feature restricts delete operations and requires extra confirmation, which helps prevent data loss and ensures cluster stability.

  1. Go to the Cluster Details page.

    1. Log on to the E-HPC console.

    2. In the left part of the top navigation bar, select a region.

    3. In the left-side navigation pane, click Cluster.

    4. On the Cluster List page, find the cluster that you want to manage and click the Cluster Name.

  2. In the left-side navigation pane, choose Nodes > Node.

  3. In the node list, select one or more target nodes.

  4. Below the list, click More, and then select Enable Deletion Protection or Disable Deletion Protection as needed.

  5. In the dialog box that appears, click OK.

Delete nodes

You can delete compute nodes that are no longer needed to scale in your cluster.

The result of deleting a node depends on its billing method. Pay-as-you-go nodes are released immediately. Subscription nodes are retained until they expire. To release a subscription node before its expiration, you must first request a refund or convert it to a pay-as-you-go node. For more information, see Release an instance.

Important

Data on a released instance cannot be recovered. To preserve the data, create a snapshot of the instance before deleting it. For more information, see Create a snapshot.

  1. Go to the Cluster Details page.

    1. Log on to the E-HPC console.

    2. In the left part of the top navigation bar, select a region.

    3. In the left-side navigation pane, click Cluster.

    4. On the Cluster List page, find the cluster that you want to manage and click the Cluster Name.

  2. In the left-side navigation pane, choose Nodes > Node.

  3. In the node list, select one or more target nodes.

  4. Below the list, click Delete.

  5. Confirm the information in the prompt, and then click OK.

Previous：OverviewNext：Manage queues