Missing configuration and report files

[sstrong99]

Description of the bug

When I run the nf-core/atacseq pipeline on NextflowTower with an aws compute environment, the only files in <outdir>/pipeline_info are samplesheet.valid.csv and software_versions.yml. The pipeline config indicates that there should be 4 other files as well:

timeline {
    enabled = true
    file    = "${params.tracedir}/execution_timeline_${trace_timestamp}.html"
}
report {
    enabled = true
    file    = "${params.tracedir}/execution_report_${trace_timestamp}.html"
}
trace {
    enabled = true
    file    = "${params.tracedir}/execution_trace_${trace_timestamp}.txt"
}
dag {
    enabled = true
    file    = "${params.tracedir}/pipeline_dag_${trace_timestamp}.html"
}

and tracedir = ${params.outdir}/pipeline_info by default.

Command used and terminal output

No response

System information

NextflowTower
aws TowerForge compute environment
Nextflow version 23.02.0-edge

[sstrong99]

FYI @maxulysse

[maxulysse]

Can confirm that I reproduce that with tower on both AWS and GCP

Tried @FriederikeHanssen trick and set up tracedir in full, and it worked...
I now have the pipeline_info folder

[mashehu]

How is that an issue with tools? Sounds more like a tower problem, or am I missing something?

[ewels]

It sounds like a config resolution issue within Nextflow to me. So likely something that needs a fix within the template.

[sstrong99]

I removed these lines from the nextflow_schema.json and that seems to have fixed it: https://github.com/nf-core/atacseq/blob/master/nextflow_schema.json#L591-L597

Maybe some kind of conflict btw the default specified there vs the one specified in nextflow.config?

[sstrong99]

and actually just deleting the default line seems to fix it as well: https://github.com/nf-core/atacseq/blob/master/nextflow_schema.json#L594

[maxulysse]

Fixed in #2317