The cluster uses a mix of shared and local storage. Knowing how to use them is essential for performance and data safety.
This page provides details of the various file systems available to users, along with their specific rules and accessibility. It is essential to understand the folder structure before using the machines.
A shared file system can be accessed from all nodes. Different usage rules and levels of accessibility apply to them all, so take your time to understand the information below. A local file system is storage that is only accessible from the machine where it is located.
| File system | Availability | Accessibility | Mount point | Backed up | Variable | Size |
|---|---|---|---|---|---|---|
| Work | Shared | Open to all users except guests | /work/<username> | No | 75TB | |
| Tools | Shared | Readable for all users | /tools | No | 10TB | |
| Home | Shared | Only to the current user | /home/<username> | Daily | $HOME | 5TB |
| Projects | Shared | Writable by group, readable by all | /projects/project_name | Daily | 50TB | |
| Seq | Shared | readable by all | /seq | Daily | 50TB | |
| Scratch | local on compute | Volatile, Accessible to jobs only | /scratch1 | No | $TEMP | 2TB |
Each user has their own /home/username folder. You cannot see other people's home folders, and vice versa (admins can if needed). Therefore, you will not be able to share files from your home folder. We kindly ask all users not to store the outputs of their analyses, raw sequencing files or software installations in the home folder. Only use it to store scripts, papers and small configuration files. Each home folder will be checked regularly, and users may find their account blocked if they do not follow these rules.
You can modify this folder from the login node; the folder will be accessible to Slurm jobs for reading only.
A daily backup will be applied to this file system.
This folder stores all the raw sequencing data after the associated metadata has been validated. Only admins can write to this folder, but all users and Slurm jobs from all nodes can view and use the data. From May 2026 onwards, the data will be stored by group.
A daily backup will be applied to this file system.
This is an open space where all users can work on their data. All new users are assigned a directory (/work//). We kindly ask all users to work on their data in this directory. There should be enough space for everyone if we all regularly sort our outputs and deletes irrelevant files. One folder per project is expected but users are free to organise this folder in any way they like.
Users with sensitive data should set the appropriate permissions for their folders, as by default every new file/folder is visible to everyone.
This file system is not backed up; therefore, we ask each user to regularly move important files in /projects (only important and relevant files). 'Not backed up' means that lost files cannot be retrieved. Therefore, it is not a good idea to leave important files here for a long time.
To ensure enough space for users over time, any files that have not been modified for over three months will be automatically deleted. A series of emails will be sent to users, 1 month, 2 weeks and 1 week before the deletion takes place. We hope that only active files and projects will be hosted here.
Users and Slurm jobs will be able to modify this file system from all nodes.
Upon request, a folder can be created for a user or group of users in this file system. This folder is intended for backing up users' final data. Users or groups of users will first be asked to provide online documentation describing the project and its files. This folder should only be used for final data that plays a critical role in the project, not for intermediate files. Users (except guests and students) can request more than one project folder and collaborate with other users within them. The intention is to have 1 folder per project and 1 documentation link associated to each project. Admins will regularly check the folder to ensure it is being used correctly. All project folders are visible to all users by default, unless users set different permissions.
Users can modify the file system from the login node, but slurm jobs only have read permissions.
This folder will be backed up daily.
Most tools will be stored in this file system. It is managed by the admins, but some users may have permission to modify it in exceptional cases. Users can load tools via the 'module' command.
This file system is only allocated when a job starts running and is deleted immediately after the job is finished or stopped. The use of it allows to speed large jobs, we recommend to use it as temporary directories during data analysis. It could be used to run an entire analysis but users should have commands to copy data back to a safe place within the job script. It is used via $TEMP variable.
Data backup is the process of creating a copy of important data on a seprate infrastructure located in a separate geographic location to ensure recovery in case of hardware failure, accidental deletion, or corruption. The backup or archive is not meant to be frequently accessed.
Users cannot directly interact with the backup, in case a file or folder needs to be restored, send an email to bioinfo.mib@wur.nl