Many HPC systems check the state of a node b efore  running a n  application, but not very many check that the performance of the node is acceptable before running the job.

Understanding inodes is key to a better understanding of HPC filesystems.

The Lustre open source distributed, parallel filesystem scales to high-performance computing environments.

IOPS is mentioned quite often when benchmarking and testing storage systems, but what does it really mean? We discuss and explain what an IOPS is, and how to measure it.

Why splat a compressed archive in your storage when you can just mount it like a storage device?

You have the cores, so use them – to compress, that is.