File Systems#

On Levante, a global Lustre file systems with a total storage capacity of 130 PetaBytes is available. Users have access to two different storage spaces on this system: WORK and SCRATCH. In addition, the HOME and SHARED storage spaces are provided by a VAST file system. Each storage space has a specific purpose as described below:

HOME is the file system where users’ sessions start upon login to Levante. It is backed up and should be used to store shell setup files, source codes, scripts, and important files. Additionally, data is stored on fast flash-backed storage. Besides backups, daily snapshots are provided. Each snapshot is kept for one week.
SHARED is a project-wide file system intended for collaborative work within a project. It should be used to store shared configuration files, source codes, scripts, documentation, and other frequently accessed small files needed by multiple project members. Data is stored on fast flash-backed storage. Daily snapshots are provided and each snapshot is kept for one week.
WORK is a project space available through the allocations process and shared between all users of a project. It provides disk space for large amounts of data, but it is not backed up. It can be used e.g. for writing raw model output and processing of data that is accessible to all project members.
SCRATCH is provided for temporary storage and processing of large data. To prevent the file system from overflowing, old data is automatically deleted. The granted retention period is 14 days.

The file systems are available on all nodes (login and compute), so you can use them during interactive sessions and in batch jobs. The table below provides further details on different storage spaces.

File system	HOME	SHARED	WORK	SCRATCH
Path	/home/[a,b,g,k,m,u]/<userid>	/shared/<projectid>	/work/<projectid>	/scratch/[a,b,g,k,m,u]/<userid>
Environment variable	$HOME
Description	Globally accessible VAST system Assigned to user account Storage of personal settings files, source codes and scripts Optimized for low latency Suitable for build and development tasks	Globally accessible VAST system Assigned to project account Storage of collaborative settings files, source codes and scripts Optimized for low latency Suitable for collaborative frequently used small data sharing	Globally accessible Lustre file system Assigned to project account Interim storage of output from running applications and frequently accessed large data Optimized for high throughput	Globally accessible Lustre file system Assigned to user account Temporary storage and processing of large data sets Regularily and indiscriminately touching all data is discouraged and will be inhibited if detected. Optimized for high throughput
Data policy:
Quota	60 GiB	500 GiB	according to annual project allocation	15 TiB
Quota check	`quo` or `quota -A -s -u $USER`	`quo` or `df -h /shared/<projectid>`	`quo` or `lfsquota.sh -p <projectid>`	`quo` or `lfsquota.sh -u $USER`
Automatic data deletion	NO	NO	NO	YES
Data life time	until user account deletion	1 month after project expiration	1 month after project expiration	14 days since the last file access
Available data recovery options:
Snapshots	YES Daily, kept for 1 week	YES Daily, kept for 1 week	NO	NO
Backup	YES please contact User Support to restore deleted files that are not covered by Snapshots	NO	NO	NO
Throughput (GB/s):
Single thread	3	3	2	2
Aggregated system peak	10	10	600	600

Snapshots on HOME and SHARED#

We provide daily snapshots that are created at 3 a.m. Each snapshot is kept for one week.

Each subdirectory in a user’s home directory, as well as each subdirectory in the /shared project directory, contains an invisible .snapshot directory. This directory has the following properties:

The .snapshot directory can only be accessed manually, for example, by running cd .snapshot.
Autocompletion is not available for this directory; you must type its name explicitly.
Once inside .snapshot, you can list all available snapshots using standard tools such as ls.
Each snapshot appears as a read-only directory, named according to the date and time it was created (UTC+2).
You can freely navigate within a snapshot just as you would in any other part of the file system.
To restore files, use standard tools like cp or rsync to copy content from a snapshot back into a writable location.

$ ls
home_2025-10-10_01_00_00_UTC  home_2025-10-12_01_00_00_UTC  home_2025-10-14_01_00_00_UTC  home_2025-10-16_01_00_00_UTC  home_2025-10-18_01_00_00_UTC  home_2025-10-20_01_00_00_UTC
home_2025-10-11_01_00_00_UTC  home_2025-10-13_01_00_00_UTC  home_2025-10-15_01_00_00_UTC  home_2025-10-17_01_00_00_UTC  home_2025-10-19_01_00_00_UTC

Finding files and disk usage on /work/#

On /work, standard find and du commands can be slow for large directories. To address this, we provide optimized commands.

Moving data on WORK#

All available Unix tools for copying and moving files on WORK can be used. However, moving files between different projects on WORK will be slow compared to moving files inside a project. Please contact us if you intend to move more than 5 TiB of data between two different projects.