Skip to main content
NetApp artificial intelligence solutions

Verify and Expand the Lustre Deployment

Contributors mcwhiteside
PDFs

Confirm cluster health, Lustre mounts, and file system capacity before production use, then use the same inventory to add building blocks when needed.

Verify the deployment

Steps

  1. From a Lustre server, confirm that all Pacemaker resources are started on the expected nodes:

    pcs status

  2. From a Lustre client, confirm that the Lustre file system is mounted:

    mount -t lustre

  3. From a Lustre client, confirm file system capacity and that MDT and OST targets are visible:

    lfs df -h

For performance validation with IOR, mdtest, and fio, see Sizing guidance.

For controlled target migration or HA failover testing, follow the procedures in the HA administration user guide.

Expand the file system

To add capacity or metadata performance, extend your existing inventory with an additional building block. Do not create a separate inventory for the new building block. The playbook automatically assigns unique MDT and OST indices and registers the new targets against the existing MGS.

Steps

  1. Add the new building block hosts, arrays, and resource groups (MDT+OST or OST-only) to the same inventory and host_vars and group_vars files you used for the original deployment.

  2. Rerun the main deployment playbook against the updated inventory.

See Solution architecture for scaling rules and Sizing guidance for guidance on when to add each building block type.