Genotoul cluster¶
Detailed information about Genotoul can be found on the Genotoul website.
Basic usage¶
Connect through SSH to the Genotoul server:
ssh thoellinger@genologin.toulouse.inra.fr
Get account infos: saccount_info thoellinger
List all available modules: module av
Load one of them: module load <name>
Some typical modules:
module load bioinfo/samtools-1.9
module load system/Anaconda3-5.2.0
Install or update a Python package for user only (does not require sudo privilege):
pip install --user <package>
To update a package for user only: pip install --user --upgrade <package>
Remember to connect to a computation node before launching any resources-demanding program:
srun --pty bash
1 cpu core and 8 gb of ram by default with srun
(2021). One may increase limits using, for instance:
srun --mem=64G --pty bash
Submitting a script using sbatch
¶
For more resources-demanding tasks or complex pipelines, it is better to write a dedicated script and to submit it to the queue using sbatch
. Example:
echo "script content" | awk 'BEGIN{print "\#\!\/bin\/sh"} {print}' > script.sh
then launch the script on slurm:
sbatch --mem=4G --cpus-per-task=1 -J <job name> --mail-user=tristan.hoellinger@inserm.fr --mail-type=END,FAIL --workdir=$PWD --export=ALL -p workq script.sh
One can visualize what is in the queue with: squeue -A thoellinger
More sophisticated examples can be found on the notebooks over here, or on Genotoul's wiki.