Closed Maarten-vd-Sande closed 4 years ago
Hello,
please include the full command line you used.
I can't reproduce the error with these arguments:
parallel-fastq-dump --split-files --gzip -t 8 -s ~/tmp/SRR2778062.1
I suspect the arguments you used were wrong because of this line:
2020-01-09T14:29:16 sra-stat.2.10.0 int: path incorrect while opening manager within database module - '/home/sande/Dropbox/Studie/PhD/snakemake-workflows/workflows/download_fastq/sra/SRR2778062/tmp'
to pass a tmp directory you need to use --tmpdir
Thanks for the reply, the full command is:
parallel-fastq-dump -s /home/sande/Dropbox/Studie/PhD/snakemake-workflows/workflows/download_fastq/sra/SRR2778062/* -O /home/sande/Dropbox/Studie/PhD/snakemake-workflows/workflows/download_fastq/sra/SRR2778062/tmp --split-spot --skip-technical --dumpbase --readids --clip --read-filter pass --defline-seq '@$ac.$si.$sg/$ri' --defline-qual '+' --threads 8 --gzip >> /home/sande/Dropbox/Studie/PhD/snakemake-workflows/workflows/download_fastq/log/sra2fastq_SE/SRR2778062.log 2>&1
(I use split spot)
It is part of a pipeline which succesfully downloaded over 300 samples already, so it would be surprising if tmpdir causes the issue. I'll try again tomorrow morning with --tmpdir
and will let you know if the problem disappears.
So this is really weird, I can't really get this to crash consistently (but it does sometimes crash) when I try it outside of the pipeline, but it always crashes inside of it. However when I do not write the output to a folder inside the folder of where our input is it seems to not crash anymore, so I guess the issue lies somewhere in that I glob all the files from this folder as input?
Anyways thanks for your reply!
yes the glob jumped out to me as well, the tmp
dir is inside the directory you are using as glob, I think this is the problem. if you want to glob a directory of SRA files make sure there isn't anything else there.
Yep that's how I "solved" it now, however still weird that it worked for tons of other samples I've run it on before.
Thanks again!
You can download the SRA here: https://sra-downloadb.be-md.ncbi.nlm.nih.gov/sos1/sra-pub-run-2/SRR2778062/SRR2778062.1
And when I dump with 8 cores it fails, normal fastq-dump performs fine
Crash happens on this line: https://github.com/rvalieris/parallel-fastq-dump/blob/fcddfa058f8e8f6ab6adffd33023d42c41205bb2/parallel-fastq-dump#L64