Closed cbmckni closed 2 months ago
I get the same error running with the locally stored CORG fastq.gz files, so that should be easier to replicate:
./nextflow -C nextflow.config kuberun systemsgenetics/gemmaker -profile k8s -v pvc-storage-test --pipeline kallisto --kallisto_index_path /workspace/projects/systemsgenetics/gemmaker/assets/demo/references/CORG.transcripts.Kallisto.indexed --input /workspace/projects/systemsgenetics/gemmaker/assets/demo/*.fastq.gz --outdir /workspace/gemmaker/output --max_cpus 4
It seems the problem is indicated with these lines in the error log:
First:
ps: error while loading shared libraries: libprocps.so.4: cannot open shared object file: No such file or directory
Followed by:
/workspace/projects/systemsgenetics/gemmaker/bin/clean_work_files.sh: line 15: perl: command not found
Then by:
stat: missing operand
I'm not sure where the ps
error is coming from because the bin/clean_work_files.sh
script which is being called doesn't use ps
. But the perl error is occurring at this line:
The clean_work_dirs
process doesn't use a container. It's just basic BASH code and since we figured perl was ubiquitous we didn't bother to create a container for this step. So, my thinking is that whatever container is being used on your k8s system is missing perl, libraries for ps
and a different stat
command.
Perhaps, the best solution would probably be to specify a container that this step should run in, that has the correct libraries, and binaries.
Going through and closing old issues. I believe this has been corrected.
Description of the bug
I am running GEMmaker on the PRP with a human kidney dataset stored locally in split fastq format. Just with the CORG reference for now to test.
Command used:
nextflow -C nextflow.config kuberun systemsgenetics/gemmaker -profile k8s -v pvc-storage-test --pipeline kallisto --kallisto_index_path /workspace/projects/systemsgenetics/gemmaker/assets/demo/references/CORG.transcripts.Kallisto.indexed --input /workspace/data/kidneydata/sra/*.fastq --outdir /workspace/gemmaker/output --max_cpus 8
nextflow.config to add k8s profile:
I get the following error, even after deleting the work dir:
Any idea what the issue may be? I have used a slight variation of this command in the past during the scidas workshop and had no issues.
Command used and terminal output