NBISweden / GenErode

GitHub repository for GenErode, a Snakemake pipeline for the analysis of whole-genome sequencing data from historical and modern samples to study patterns of genome erosion.
GNU General Public License v3.0
23 stars 7 forks source link

Update cluster.yaml #33

Closed verku closed 1 year ago

verku commented 1 year ago

Users reported:

"At least two rules on the culster.yaml file fastqc_historical_merged (named the same in the rule's file "1.1_fastq_processing.smk") and convert_historical_sam (named as sai2bam in the rule's file "2_mapping.smk" ) are not getting assigned the resources (time and cpus-per-task) written for the specific rule and instead are just getting assigned the minimum default parameters defined at the beginning of the cluster.yaml. As a consequence of this bug, jobs keep getting killed because of OUT OF TIME. In the case of fastqc_historical_merged, even though both in the cluster.yaml and in the rule's file ask for 2 cores, it only gets assigned 1. The only solution we found was to change the default from 1 to 2. In the case of sai2bam is a bit trickier since the rule asks for 8 cores and 10 days and is only getting assigned 2 cores and 2 hours. The temporary solution I find here is to extend the time set in the default parameters from 2 to 4 or 6 hours (as I don't think it is a good idea to keep increasing the default minimum number of cores). Seems like Snakemake is having trouble passing the right instructions onto Slurm in these two specific cases. "

"Changing the rule name in the cluster file from convert_historical_sam to sai2bam should do the trick. If things have different names in the cluster and rule file, snakemake can’t link them to each other. "