Lustre 101

These slides were originally hosted by ORNL in 2016 as part of a self-paced web-based course series designed to provide an introduction to Lustre. The course was targeted at experienced system administrators who were relatively new to Lustre. Although the content may be old, some of information/concepts presented in this course are still relevant to newer versions of Lustre.

• Introduction to Lustre - This presentation provides a general overview of the Lustre file system for anyone wanting to learn more about basic Lustre functionality, features, and architecture. The basic components of Lustre are discussed, including the LNet transport layer. Information about Lustre file striping is also included.
• Hardware Selection and Benchmarking for Lustre
• Basic Lustre Installation and Setup from Stock RPMs
• Creating a Lustre Test System from Source with Virtual Machines
• Lustre Tuning and Advanced LNet Configuration
• File System Administration and Monitoring
• Analysis of Crash Dumps and Log Files
• Evaluating the Functionality, Performance, and Reliability of Lustre
• LNet Bonding
• Lustre and Memory
• Lustre Over Long-Haul Connections Using LNet Routers
• Working with Problematic Nodes
• Recovery and Eviction

The following tutorials were presented at the International Workshop on the Lustre Ecosystem held in 2015 and 2016:

2015

• Network Contention and Congestion Control: Lustre Fine-Grained Routing
• Improve Large-Scale Storage System Performance via Topology-aware and Balanced Data Placement
• LNET Configuration
• Robust Health Monitoring
• Failure Handling
• An Introduction to fileUtils

2016

• A 12 Step Program for Lustre Filesystem Addiction
• Lustre Job Stats Metric Aggregation at OLCF
• Managing Lustre on a Budget
• Securing Lustre with Nodemap and Shared Key
• How to Install, Tune, and Monitor a ZFS Based Lustre File System