These are chat archives for nextflow-io/nextflow

21st
Aug 2018
Ashley S Doane
@DoaneAS
Aug 21 2018 01:14

when specifying a conda environment and sge executor, I get an error :

cat work/a6/c9eea2ea1f3a5725e3b9a6c0b2e548/.command.log

/opt/sge/default/spool/execd/node161/job_scripts/1749279: line 68: activate: No such file or directory

the conda env works when using local executor. Any ideas what might be going on? If I qlogin to a compute node, source activate <env> works.

Guillaume Collet
@gcollet
Aug 21 2018 07:45
Hi, I was trying to loop between processes but it is impossible with nextflow, isn't it ?
It must be a DAG
Maxime Garcia
@MaxUlysse
Aug 21 2018 07:51
No, you can do it
Guillaume Collet
@gcollet
Aug 21 2018 08:27
Great ! thanks @MaxUlysse
Maxime Garcia
@MaxUlysse
Aug 21 2018 08:34
You're welcome
Jose Espinosa-Carrasco
@JoseEspinosa
Aug 21 2018 09:38

hey guys!
In my hands it seems that when I resume a pipeline I can not modify the number of queues I want to use (process.queueSize) editing the nextflow.config:

manifest {
    mainScript = 'annotate-mutations-simulated-wgs.nf'
}

profiles {
    standard {
    process.executor = 'local'
    }

    cluster {
    process.executor = 'slurm'
        process.queueSize = 1
        process.cpus = 5
        process.memory = '1 GB'        

    }
}

As you see I want to reduce the queueSizeto be 1. The reason is that I just want to run a test without occupying additional nodes in the cluser. Am I doing something wrong or is this behavior expected?

Maxime Garcia
@MaxUlysse
Aug 21 2018 09:42
I think that queueSize in on the executor scope
Jose Espinosa-Carrasco
@JoseEspinosa
Aug 21 2018 09:44
Thanks!
Sander Bervoets
@Biocentric
Aug 21 2018 17:38
@tbugfinder Concerning my segfaults in Salmon quant, no changing image, instance and increasing memory did not help
Paolo Di Tommaso
@pditommaso
Aug 21 2018 18:50
which base image are you using? the same container is working locally? likely this is a salmon problem.