At the end of the nice wrapper script run_pipeline.sh, the pipeline is submitted as job to a node in the slurm cluster. In my case, this failed since the queue computepart was still hardcoded in the script. Here it would be nice to have an additional argument for the wrapper script that allows you to specify the queue (and maybe use default queue if the argument is not set).
Furthermore it would be nice if one could also run the pipeline entirely on the head node. In our cluster architecture, only the head/login node is allowed to submit jobs. This means that you would need to run the pipeline on this node. But I know that this is a bit unusal.
At the end of the nice wrapper script
run_pipeline.sh, the pipeline is submitted as job to a node in the slurm cluster. In my case, this failed since the queuecomputepartwas still hardcoded in the script. Here it would be nice to have an additional argument for the wrapper script that allows you to specify the queue (and maybe use default queue if the argument is not set).Furthermore it would be nice if one could also run the pipeline entirely on the head node. In our cluster architecture, only the head/login node is allowed to submit jobs. This means that you would need to run the pipeline on this node. But I know that this is a bit unusal.
Andreas