Closed TBradley27 closed 10 months ago
@TBradley27 we try not to set any hard resource requirements on the processes to allow flexibility. Despite this, most distributed system will have default resource limits that might be too low for the workflow to run successfully and over which we have no control. For these use cases, you will to discuss with your IT provider to help you define a configuration file that works fine for your environment. You might also find help by having a look at the custom configurations provided by nf-core, which might provide defaults that would work fine with your system.
We thank for the feedback, we will update the documentation to help users avoid coming across these issues in the future :)
Operating System
CentOS 7
Other Linux
No response
Workflow Version
v0.3.0
Workflow Execution
Command line
EPI2ME Version
No response
CLI command run
srun bash /scratchb/jblab/ampliconseq/nextflow run epi2me-labs/wf-somatic-variation \ -resume \ -c my_config.cfg \ -profile 'singularity' \ --snv \ --sv \ --mod \ --sample_name 'XXXX' \ --ref 'hsa.GRCh37_g1kp2.fa' \ --bam_normal 'XXXXXXX.sorted.bam' \ --bam_tumor 'XXXXXXX.sorted.bam'
Workflow Execution - CLI Execution Profile
custom
What happened?
Hello,
I was trying to use this workflow with an adaptation of the singularity profile. I adapted the singularity profile like this so singularity containers could be deployed using SLURM job scheduling software:
The problem that I encountered is that the default resource and memory allocations for each job was too low. By default each job was only allocated 1GB of memory and 1 hour of wall clock time. I have standard ~30X coverage for my normal sample and ~30X for my tumour sample. Some jobs such as
somatic_sv:nanomonsv_get
have a standard run time of over four hours. I also think it likely that the default allocation of 1GB produces memory bottle necks too. From my report,somatic_sv:nanomonsv_get
used 25GB of virtual memory.snv:clairs_full_hap_filter
also did not complete within one hourI resolved this issue by editing the config file and editing the process section like so to increase memory and time allocations for each process:
Relevant log output
Application activity log entry
No response