On 10/4/22 21:41, Nagle, Michael F wrote:
Hello,
Thanks in advance to the developers and community. I'm a biologist who
is not an expert in Linux and am dealing with a pretty rudimentary
issue I haven't been able to solve.
In short, I need to figure out the correct way to change
$Global::max_file_length to avoid an error resulting from output
folder path name being too long, or to figure out a simple workaround.
_/When I launch `parallel` to run over some jobs.../_
parallel --results output -a $job_list_name
_/I get:/_
mkdir output/1/gemma -bfile .\_1323_cohort_maf01_geno10.snp.pass -p
.\_pheno_files\_aaaaatest.noheader.pheno -k .\_1323_cohort_maf01_geno10.cXX.txt
-lmm 1 -n 3 -miss 0.1 -o testa1: Invalid argument at
/home/gmobot/anaconda3/envs/gemma/bin/parallel line 12170.
_/Looking at line 12170, I see:/_
12163 # mkpath failed: Argument probably too long.
12164 # Set $Global::max_file_length, which will keep the individual
12165 # dir names shorter than the max length
I have tried changing max_file_length to all sorts of values, from 40 to 4000,
by the two methods:
1. max_file_length=40
2. export max_file_length=4000
I then echo'd $max_file_length to make sure it was set as intended before
running parallel again. I'm still getting the same issue.
This issue is happening with both the Biden and Elizabeth releases of GNU
Parallel, installed via conda. It did not happen when running the same code on
the same computer with Ubuntu 20.04 LTS, but is now happening with Ubuntu 22.04
LTS.
What am I doing wrong? What's the correct way to set the max_file_length global
variable, and what should I set it to so I can avoid this issue?
As an alternative solution, is there a simple way for me to run GNU parallel
that doesn't create a folder based on the names of my (very long) commands? I
suppose I could produce a separate wrapper script for each command and pass
that, but would prefer a more elegant and less tedious solution.
Thanks!
Your "I get:" bit is using backslashes. Is that from you or the parallel?
