We got this error

Cannot invoke "java.nio.file.Path.getFileSystem()" because "path" is null

running nf-core/magmap on

dardel

nextflow run nf-core/magmap -r dev -profile pdc_kth -params-file params.yml -c nextflow.log

Nextflow version: 25.04.7 the pipeline was launched on dardel. my run worked fine but maybe @Nicolo you can give more details. (write here content of

params.yml

and

nextflow

.config maybe)

Hello, params.yml: project: naiss2025-5- outdir: results_magmap input: fastq_samplesheet.csv genomeinfo: MAG_samplesheet.csv skip_qc: true skip_fastqc: true skip_trimming: true skip_sourmash: false sourmash_save_sourmash: true sourmash_save_prefetch: true sourmash_save_prefetch_csv: true bbmap_save_bam: true bbmap_save_index: true save_concatenated_genomes: true nextflow.config: process { withName: BBMAP_BBDUK { memory = 128.GB cpus = 20 time = '1d' } withName: BBMAP_INDEX { memory = 1.TB cpus = 64 time = '2d' } withName: PROKKAGFF2TSV { memory = 400.GB cpus = 24 time = '1d' } withName: BBMAP_ALIGN { memory = 1.TB cpus = 128 maxForks = 5 time = '3d' } } running on dardel nextflow run nf-core/magmap -r dev -profile pdc_kth -params-file params.yml -c nextflow.config Error: ERROR ~ Cannot invoke "java.nio.file.Path.getFileSystem()" because "path" is null -- Check script '/cfs/klemming/home/n/nicoivan/.nextflow/assets/nf-core/magmap/subworkflows/nf-core/utils_nfcore_pipeline/main.nf' at line: 82 or see '.nextflow.log' file for more details ERROR ~ Pipeline failed. Please refer to troubleshooting docs: https://nf-co.re/docs/usage/troubleshooting -- Check '.nextflow.log' file for details -[nf-core/magmap] Pipeline completed with errors

My hunch is that there's an output file missing from one of the processes. Can you check in

.nextflow.log

which was the last process to run? Starting from there, I'd check if unsetting one or more of the

save_

params might help. (I never run with them, and it never failed. OTOH, missing files should be detected even if

save_

something is not set, so I'm looking for a straw here...)

Here is where the error appeared: Nov-11 100103.386 [main] DEBUG nextflow.script.ScriptRunner - > Awaiting termination Nov-11 100103.386 [main] DEBUG nextflow.Session - Session await Nov-11 100103.387 [Actor Thread 250] ERROR nextflow.extension.OperatorImpl - @unknown java.lang.NullPointerException: Cannot invoke "java.nio.file.Path.getFileSystem()" because "path" is null Nov-11 100103.404 [Actor Thread 250] DEBUG nextflow.Session - Session aborted -- Cause: Cannot invoke "java.nio.file.Path.getFileSystem()" because "path" is null

What I'm after is the process name that seem to be the last executed. Process names look like

BBMAP_ALIGN

. You might need to go further up in the log to see what Nextflow was waiting for. Or, if you still have the session output you could perhaps see which process(es) were running at the time the pipeline failed.

This might be better

As far as I can see, none of the processes started. Did this happen very early? If so, which Java are you using? I use

java/17.0.4

, i.e. like this:

module load bioinfo-tools PDC apptainer java/17.0.4

.

It happens almost immediately, right after listing all the steps. I wasn't manually loading java but now I've loaded java/17.0.4 and removed all save_ params but still got the same error

Are you using your own Nextflow or the module? (I use my own, which is why I need to load the java module.)

I'm using the module. should I try create a conda env and install all dependencies?

Unload the Nextflow module and install Nextflow in your

~/bin

directory:

cd ~/bin; curl -s <https://get.nextflow.io> | bash

(assuming your

~/bin

is in your

PATH

).

I could reproduce the same error on my local computer running

nextflow run nf-core/magmap -r dev -profile test,docker -params-file params.yml

if I remove from the

params.yml

it works but then it fails again on another bug I found. so that's good to know!

Just tried removing: skip_qc: true skip_fastqc: true skip_trimming: true as Danilo suggested and it works

So, it's skipping the trimming and qc that fails? Should be easy to fix, great that you found this @Danilo Di Leo.

there are two different bugs. 1. skip_trimming + bbduk 2. skip_fastq

true

or skip_qc

true

OK. Probably related somehow as it's all preprocessing.