Discovery Emergency Maintenance 5/27 and 5/28

Dear Discovery testers,
I’d like to let you know that we are diligently working on the current issue with the Lustre file system. We also wanted to take this chance to do our scheduled maintenance instead of taking the system down again on 27th and 28th. The Discovery system is going to be down a bit longer, but will come with more nodes and more stable file system. During the maintenance period, we are going to work on:

  • Adding 20 more compute nodes
  • Updating system monitoring tools
  • Implementing system regression test process
  • Fixing Lustre file system issue

We will keep you updated on this and will do our best to bring the system back ASAP. We appreciate your patience and understanding.

Thank you, and I hope you and your household are safe and well.


Byoung-Do (BD) Kim, PhD
Director, USC Center for High-Performance Computing
Adjunct Professor of Data Science, Marshall School of Business