Skip to content

Disk Space and Storage Options

Main Storage Disks

/softdec

  • Size: 100 GB
  • Configuration: Part of a larger 300 TB disk, not saved, under RAID 4.
  • Purpose: This partition is dedicated to general software that is used across various projects. It is not suitable for personal data storage.
  • Example: Store shared applications and tools that need to be accessed by all users.

/datadec

  • Size: 300 TB
  • Configuration: Part of the primary 300 TB disk, not saved, under RAID 4.
  • Purpose: Designed for storing your active project data.
  • Example: Use this space for ongoing projects where data accessibility and speed are crucial.

Additional Disks

/roofdec

  • Size: 100 TB
  • Purpose: This disk is intended for data storage related to active projects and collaborations.
  • Example: Ideal for storing large datasets that are frequently accessed and modified during the course of your research.

/loftdec

  • Size: 100 TB
  • Purpose: Allocated for archiving purposes.
  • Example: Use this space to store completed project data or historical data that is accessed infrequently.

Local Storage on Computing Nodes

/data

  • Size: 860 GB
  • Configuration: Local disk, not saved, not under RAID.
  • Purpose: This disk is recommended for storing temporary files that are specific to individual jobs or sessions.
  • Example: Temporarily store intermediate computation results, logs, or files that are required for running short-term processes.

Distributed File System (GlusterFS)

/scratch

  • Size: 7 TB
  • Configuration: Distributed file system, not saved, not under RAID, replicated across 3 nodes.
  • Purpose: Enables shared access to data across different nodes, facilitating collaboration and data sharing among various jobs.
  • Example: Store data that needs to be accessed or modified by multiple jobs running on different nodes, such as input files for simulations that are processed in parallel.
  • Note: GlusterFS uses the space available on /data disks of individual nodes, making it accessible across the cluster. This shared space is ideal for temporary data needed during job runs but should not be used for long-term storage due to its non-redundant configuration.