Skip to main content
NetApp artificial intelligence solutions

Deploy Lustre with NetApp E-Series Storage

Deploy Lustre with NetApp E-Series Storage by completing the hardware, environment preparation, inventory, Ansible deployment, and verification tasks in order.

Prerequisites

Before you start the deployment, ensure the following:

  • You followed the sizing and building-block guidance in Sizing guidance and selected hardware that matches Hardware components.

  • You confirmed that the server, HCA, E-Series, operating system, Lustre, SANtricity OS, and DOCA-OFED combination is listed for E-Series Lustre in the NetApp Interoperability Matrix Tool.

  • All OSS/MDS server nodes and E-Series arrays for the planned building blocks are on site and ready to rack.

  • You have credentials and network access for server BMCs, array management ports, and the host that will serve as the Ansible control node.

Note NetApp supports deploying the Lustre HA cluster and Lustre clients with the NetApp E-Series Lustre Ansible collection. Do not configure Lustre targets, LNet, NVMe-oF host connectivity, or Pacemaker resources manually.

Deployment workflow

Complete the following tasks in order:

  1. Rack and cable Lustre hardware.

    Rack the Lustre servers and EF80 arrays, then cable the backend NVMe-oF and frontend LNet fabrics.

  2. Prepare the Lustre deployment environment.

    Configure the arrays and servers, then install Ansible and its dependencies on the control node.

  3. Customize the Lustre Ansible inventory.

    Select the template for the site's fabric and populate the servers, arrays, networks, building blocks, and credentials.

  4. Deploy the Lustre HA cluster and clients.

    Prepare the server operating system, install NVIDIA DOCA-OFED, run the main deployment playbook, and configure Lustre clients.

  5. Verify and expand the Lustre deployment.

    Confirm cluster health, mounts, and capacity before production use. Extend the existing inventory when the file system needs another building block.