Skip to main content
AI Data Engine

Add data compute nodes to your AI Data Engine cluster

Contributors netapp-dbagwell dmp-netapp
PDFs

You can add data compute nodes as part of creating a new AI Data Engine (AIDE) cluster or expanding an existing cluster. The workflow consists of discovering and configuring the nodes using ONTAP System Manager.

Prepare to add nodes

Consider the following when adding data compute nodes.

Minimum number of data compute nodes

Exactly three available data compute nodes are required for a new AIDE cluster.

Hardware installation and addressability

Make sure the following prerequisites are met:

  • The new data compute hardware is racked, powered on, and cabled to the cluster switches.

  • You have an IP address space range available for the data compute nodes to ONTAP backend subnet.

  • The ONTAP cluster has been initialized and is reachable from the cluster management LIF.

System Manager credentials

You need storage administrator privileges to perform AIDE cluster creation or expansion tasks.

Software compatibility

Review the following documentation to confirm that your data compute hardware and software versions are compatible with your ONTAP cluster:

During the node add operation, System Manager will confirm the new nodes run a software version compatible with:

  • The ONTAP cluster effective version if this is the first data compute node added.

  • The existing data compute cluster version if data compute nodes are already present.

If a node is incompatible:

  • An error is shown next to the affected data compute node in the Add dialog.

  • You must first update the data compute node software (or ONTAP, as appropriate) to a compatible version.

Add the data compute nodes

Add data compute nodes as part of creating a new AIDE cluster or expanding an existing cluster.

Steps

  1. In System Manager, select Dashboard in the navigation pane and then the Health card.

  2. Confirm that there are nodes to add and select View details to display the list.

    The list contains discovered nodes that are not yet part of the AIDE cluster.

  3. Alternatively, you can select Cluster and Overview and the Data compute tab to see the list.

  4. At the bottom of the data compute page, select Add above the list of nodes.

  5. In the Add nodes dialog, select the data compute nodes you want to add.

  6. If this is the first time you're adding nodes and no backend subnet exists, select Add subnet and provide:

    • Subnet name (for internal use)

    • Subnet address and mask

    • IP address range for data compute nodes and ONTAP nodes on the backend network

      System Manager validates that the range includes enough free IP addresses for all data compute nodes being added, all ONTAP nodes in the cluster, and additional cluster-level floating IPs used for data compute node to ONTAP communication.

  7. Do the following regardless of whether you added the backend subnet or it already exists:

    1. Review the available IP addresses.

    2. If needed, select Edit subnet and extend the IP range.

      • You can only grow the range. Shrinking or changing the subnet is not supported.

      • Changing the subnet or IP range might require recreating the underlying Kubernetes cluster on data compute nodes and can take several minutes.

  8. Optionally configure the Data Engine service interface by providing:

    • Service IP address

    • Network mask

    • Gateway (if required for your environment)

      The single IP will be load-balanced across data compute nodes and used as the frontend address for AIDE Console and related APIs.

  9. Review the selected nodes, backend subnet, and Data Engine service interface settings.

  10. Select Add and wait for the operation to complete. System Manager will perform the following actions:

    • Add the selected nodes to the data compute cluster

    • Provision the backend networking and join nodes to the Kubernetes-based data compute cluster

    • Update internal metadata for data compute node discovery

  11. After completion, select Cluster and Overview and confirm:

    • Under Data compute the new data compute nodes appear online as part of the cluster

    • All nodes are Healthy

    • AIDE version is displayed correctly for each node

    • Verify the dashboard Health card shows the updated node count

  12. (Optional) After all nodes join the cluster, verify that all services start correctly.

    1. Verify Kubernetes pod status:

      kubectl get pods -n aide-system

    2. Confirm all pods show Running or Completed status.