Identifying memory leaks in nextflow workflow

[anoronh4]

We are looking to make the main nextflow run more memory efficient by identifying memory leaks and objects that require a lot of the heap. However we are unsure of where to start and have not found a lot of guidance on how to best address it specifically for nextflow. We are processing >10,000 and at one process, a sample will spawn about 60 jobs for variant calling different parts of the genome. this means our runs are submitting a lot of jobs and we also have a large amount of channels that don't close because we are using watchPath on up to 4 files at a time. the maps are few, quite small and updated infrequently.

We are already fine-tuning NXF_OPTS and heap size variables, but this seems to just be postponing the problem to another day instead of addressing underlying issues in the workflow. any guidance would be appreciated.

finally i do want to mention that i have a TimerTask in my workflow:

def touchInputs() {
  new Timer().schedule({
  for ( i in epochMap.keySet() ){
    fileEpoch = file(i).lastModified()
    if ( fileEpoch > epochMap[i]) {
      epochMap[i] = fileEpoch
      "touch -ca ${i}".execute()
    }
  }
} as TimerTask, 1000, params.touchInputsInterval * 60 * 1000 ) // convert minutes to milliseconds
}

the map operated on in the above snippet has only four keys in it and i do not believe it to be a memory leak, but maybe someone else would be able to identify something. touchInputs() is only run once.

[pditommaso]

Too broad context. You can use tools and techniques for Java application to try identifying the source of the leak.