Closed cement-head closed 3 years ago
Hi @cement-head ,
this looks like an issue with the linux version (what OS are you using?). Your system most likely does not support the (safer) -o pipefail option (https://vaneyckt.io/posts/safer_bash_scripts_with_set_euxo_pipefail/). You are probably good removing it from all the scripts where it is present.
Ubuntu 18.04.5 LTS; okay, so that would mean to delete the -o option at the beginning of each script?
Example:
Change this ORIGINAL:
#!/bin/bash
set -e -o pipefail
#++++ This script is part of: ++++
#++++ mitoVGP v2.2 ++++
#++++ The Vertebrate Genomes Project Mitogenome Assembly Pipeline ++++
#++++ Credit: Giulio Formenti gformenti@rockefeller.edu ++++
TO this?
#!/bin/bash
set -e pipefail
#++++ This script is part of: ++++
#++++ mitoVGP v2.2 ++++
#++++ The Vertebrate Genomes Project Mitogenome Assembly Pipeline ++++
#++++ Credit: Giulio Formenti gformenti@rockefeller.edu ++++
Also, I'm a little unclear as to how to specify the input reads file (list). Would using the -1 option and giving the script the fastq filename be appropriate?
Example:
$ ./mitoVGP -a pacbio -s Rana_sylvatica -i aRanSyl -r NC_027236.1.fasta -t 124 -1 wfpbs2-abcd.fastq
Okay, dash in Ubuntu (not bash) doesn't recognise pipefail: https://askubuntu.com/questions/886537/set-e-o-pipefail-not-working-on-bash-script-on-ubuntu-16
Thanks for checking this. It then seems more like a symlink issue in your environment, since the shebang line requests bash (#!/bin/bash
). If you confirm we can close I think. Thanks!
Yes; switching to bash prompt doesn't help. I'm using a very vanilla 18.04.5 LTS install; just for bioinformatics. Looks as though starting from bash bounces it back to dash. Current workaround is to comment out
#set -e -o pipefail
Thanks, good for other users who may encounter the same issue to know..closing!
What DISTRO was this script coded on? CentOS? Other RedHat derivative? We're running into other problems; looks as if this (these) scripts are DISTRO specific.
Currently working fine on Red Hat Enterprise Linux Server 7.5. Tested a while ago on CentOS also worked fine. Maybe open a new thread with the specific issues?
Crashes on start: