fsx.l.wrkshp.2018.11
© 2018 Amazon Web Services, Inc. and its affiliates. All rights reserved. This work may not be reproduced or redistributed, in whole or in part, without prior written permission from Amazon Web Services, Inc. Commercial copying, lending, or selling is prohibited.
This will show solutions architects how to take advantage of a fully-managed Lustre file system with Amazon S3 integration that links the file system to an existing data repository.
This workshop will cover how to create a file system and load metadata from an Amazon S3 bucket. Learn how to lazy load objects from Amazon S3 to an FSx for Lustre file system or review some sample scripts to bulk load all objects with a single operation. FSx for Lustre file systems are designed for compute-intensive workloads and can achieve high levels of throughput and operations per second with consistent low latencies. Test read and write performance as you walk through these step-by-step instructions and monitor how hard your tests are driving the file system by creating an Amazon CloudWatch dashboard.
You will be using your own AWS account for all workshop activities. WARNING!! This workshop will exceed your free-usage tier. You will incur charges as a result of stepping through this workshop. Terminate all resources that were created during this workshop so you don’t continue to incur ongoing charges.
Click on the link below to go to the Amazon FSx for Lustre prerequisites section. It is recommended to complete this section before moving to The Workshop.
Step | Workshop |
---|---|
0 | Prerequisites |
Click on the links below to go to the FSx for Lustre workshop. You must first complete Prerequisites above before moving on to the steps in The Workshop below. These steps must be completed in order.
Step | Workshop |
---|---|
1 | Create a file system |
2 | Create a CloudWatch dashboard |
3 | Launch a client |
4 | Mount the file system |
5 | Examine the file system |
6 | Lazy load files |
7 | Bulk load files |
8 | Performance tests |
This sample code is made available under a modified MIT license. See the LICENSE file.