Skip to content

Overview

Your journey on the ULHPC facility is illustrated in the below figure.

In particular, once connected, you have access to several different File Systems (FS) which are configured for different purposes.

What is a File System (FS) ?

A File System (FS) is just the logical manner to store, organize & access data. There are different types of file systems available nowadays:

  • (local) Disk FS you find on laptops and servers: FAT32, NTFS, HFS+, ext4, {x,z,btr}fs...
  • Networked FS, such as NFS, CIFS/SMB, AFP, allowing to access a remote storage system as a NAS (Network Attached Storage)
  • Parallel and Distributed FS: such as SpectrumScale/GPFS or Lustre. Those are typical FileSystems you meet on HPC or HTC (High Throughput Computing) facility as they exhibit several unique capabilities:
    • data is spread across multiple storage nodes for redundancy and performance.
    • the global capacity AND the global performance levels are increased with every systems added to the storage infrastructure.

Storage Systems Overview

Current statistics of the available filesystems are depicted on the side figure. The ULHPC facility relies on 2 types of Distributed/Parallel File Systems to deliver high-performant Data storage at a BigData scale:

  • IBM Spectrum Scale, formerly known as the General Parallel File System (GPFS), a global high-performance clustered file system hosting your $HOME and projects data.
  • Lustre, an open-source, parallel file system dedicated to large, local, parallel scratch storage.

In addition, the following file-systems complete the ULHPC storage infrastructure:

  • OneFS, A global low-performance Dell/EMC Isilon solution used to host project data, and serve for backup and archival purposes
  • The ULHPC team relies on other filesystems within its internal backup infrastructure, such as xfs, a high-performant disk file-system deployed on storage/backup servers.

Summary

Several File Systems co-exist on the ULHPC facility and are configured for different purposes. Each servers and computational resources has access to at least three different file systems with different levels of performance, permanence and available space summarized below

Directory Env. file system backup
/home/users/<login> $HOME GPFS/Spectrumscale no
/work/projects/<name> - GPFS/Spectrumscale yes (partial, backup subdirectory)
/scratch/users/<login> $SCRATCH Lustre no
/mnt/isilon/projects/<name> - OneFS yes (live sync and snapshots)

Last update: September 14, 2024