We have completed the platform upgrade and migrated to AlmaLinux 9.6. In this regard, we would like to draw your attention to several changes:

1. New queues for HybriLIT and Govorun;
2. The Lmod module system is now used instead of Modules;
3. User home directories have changed:
   • /zfs/store5.hydra.local/user/l/login — the user home directory, intended for file editing, building, and compiling programs.
   • /lustre/projects/l/login — the user directory for running computations, data storage, and data processing.
4. The recommended workflow for working on the platform is described here.

We also inform you that file copying from the previous storage to your current directory /lustre/projects/l/login is still in progress. This process may take some time. Please do not worry — all your data are safe and will be fully transferred. In addition, updated versions of application software packages have been installed on the platform. If you require additional software, please submit a request here.

Recommended Steps for Working on the HybriLIT Platform

A typical workflow on the platform consists of several steps: connecting to the cluster, preparing files in your home directory, copying data to the project filesystem for computations, running jobs via SLURM, and saving results.

Workflow Steps

1. Connecting to the Platform
   Connect to the cluster via SSH:
   ssh hydra.jinr.ru -l <login>

   After logging in, you will be in your home directory.

2. Working in the Home Directory (HOME, ZFS)
   Path: /zfs/store5.hydra.local/user/l/login/

   Purpose:

   The home directory is intended for:
   • Editing source files
   • Building and compiling programs
   • Storing configuration files
   • Preparing scripts for job submission
   

   Quota: 100 GB

3. Loading Software via Lmod
   The platform uses the Lmod module system to manage software.
   
   To see available programs: module avail
   To load a module: module load <module_name>
   
   

   To load a module: module load <module_name>

   
   Example:
   module load gcc
module load openmpi

   
   After loading the necessary modules, you can compile programs or prepare the environment for job execution.

4. Copying Files to the Project Filesystem (PROJECTS, Lustre)
   Before running computations, it is recommended to copy data and scripts to the project filesystem: /lustre/projects/l/login/
   
   

   Purpose: This filesystem is intended for:

   • Job input data
   • Intermediate files
   • Computation results

   It is optimized for parallel I/O and running computational tasks.

    

   Quota: No quota is set for users.

   Example of copying files: cp input.dat /lustre/projects/l/login/
   

5. Running Jobs via SLURM
   To run computations, prepare a SLURM job script and submit it to the queue.
   
   Submitting a job: sbatch job.sh

   The SLURM scheduler will allocate the job to the compute nodes of the cluster.
   

6. Retrieving Results
   After the job completes, an output file will appear in the working directory: slurm-<jobid>.out

   All files created by your program will also be located in the project directory.

7. Copying Results Back to HOME
   Important results should be copied back to your home directory for storage or further processing:
   cp results.dat /zfs/store5.hydra.local/user/l/login/