- Add your account name for Hilbert to:
profile/hilbert_cluster.json - You can specify the location to write stderr and stdout output to in
profile/hilbert_cluster.json(make sure that the folders exist already!). In this example, they are written to log/cluster_jobs/stdout/ and log/cluster_jobs/stderr/ - Snakemake specific parameters can be set in:
profile/config.yaml
You can use any Snakemake workflow, just make sure that you add "resources" to each rule that needs more resources than the default specified in profile/config.yaml (see example Snakefile).
-
Install snakemake via conda (I always have problems using snakemake available via "module load"). Details on how to use conda on Hilbert are explained here: https://wiki.hhu.de/display/HPC/Conda
-
log in to the login node
-
go to the snakemake node:
ssh snakemake-node -
open a screen session:
screen -
activate conda environment that has snakemake installed in it
-
run pipeline:
snakemake --profile profile --cluster-status ./hhu_hilbert.py
You can monitor the jobs created by the pipeline on myjam: myjam3.uni-duesseldorf.de/