Direct connection to compute nodes is forbidden in the Pyrene cluster. The only way to access resources allocated to a job is to submit another job. To do this, use the `srun` command to connect inside the targeted job. Assuming you have submitted a _job_id_ batch job:
```
$ sbatch script.sh
Submitted batch job job_id
```
simply use the _job_id_ allocated by Slurm to connect to the corresponding node: