GitHub
01/06/2025, 12:32 PM$SLURM_ACCOUNT is used as the --account. Nextflow submissions will fail because the dedicated nodes have different --account requirements.
This PR has changes to dynamically evaluate whether the outcome of the queue variable is one of the dedicated nodes and will give change the --account cluster option based on this. Another option is to have users specify multiple environment variables and call these instead of the hardcoded --account in clusterOptions like I have done now.
Since there are issues with evaluating queue, I repeated the queue definition statement. It's a bit ugly so if you have other suggestions, feel free.
I have only tested this config for the dedicated vs non-dedicated CPU nodes, not for the gpu or the gpu_h100.
nf-core/configs
✅ All checks have passed
30/30 successful checksGitHub
01/10/2025, 9:47 AM