Getting Started with Amazon FSx for Lustre

Verify that your environment meets the system requirements, complete the file server service catalog, add a subclient, and then perform a test backup and restore.

Procedure

  1. Verify that your environment meets the requirements.

    • Access nodes for Amazon FSx for Lustre must be running Red Hat-based Linux Distributions supported by Amazon.

    • Lustre is supported only on x64 processors.

    • The fileset feature is not supported.

    • There is no access to MDS/MDT servers for optimized scanning.

      • Each 100 million files require approximately 3 hours and 10 minutes of scan time, and every 100 to 150 TB of application data needs one large access node.

      • Commvault recommends having a minimum of two access nodes for improved parallelization and redundancy.

      • Load balancing across nodes gives better overall performance.

      Example Scan Time

      Items scanned Time taken to scan Content
      100 million in 1 subclient 3 hours 10 minutes Entire volume: /lustrefs

      Note

      For faster scanning times with a very large number of files, especially where there is no access to the MDS for fast incremental scanning, you can create multiple subclients and split the data across them to run multiple jobs in parallel.

  2. Add Amazon FSx for Lustre file storage.

  3. Perform a backup.

  4. Perform a restore.

×

Loading...