Segmentation fault with STAR container

[valleema]

Check Documentation

I have checked the following places for your error:

• nf-core website: troubleshooting
• nf-core/rnaseq pipeline documentation

Description of the bug

Arriving at the STAR_ALIGN process, I have a Segmentation fault, with nothing else useful outputted.

Steps to reproduce

Steps to reproduce the behaviour:

As used in:

rnaseq/modules/nf-core/modules/star/align/main.nf

Line 17 in 8094c42

container 'https://depot.galaxyproject.org/singularity/star:2.6.1d--0'

1. Command line:
   singularity pull --name star-2.6.1d--0.img https://depot.galaxyproject.org/singularity/star:2.6.1d--0
singularity exec star-2.6.1d--0.img STAR --version
2. See error:
   Segmentation fault

Expected behaviour

With a more recent STAR:

1. Command line:
   singularity pull --name star-2.7.9a--h9ee0642_0.img https://depot.galaxyproject.org/singularity/star:2.7.9a--h9ee0642_0
singularity exec star-2.7.9a--h9ee0642_0.img STAR --version
2. See output:
   2.7.9a

Log files

Have you provided the following extra information/files:

• The command used to run the pipeline nextflow run nf-core/rnaseq -profile singularity --genome GRCh38 --max_cpus 18 --max_memory 128.GB -resume --input samplesheet.csv -ansi-log false
• The .nextflow.log file: nextflow.log

System

• Hardware: Local server
• Executor: No executor, local
• OS: Debian
• Version 11

Nextflow Installation

• Version: 21.04.1

Container engine

• Engine: Singularity
• version: 3.8.0

[valleema]

OK I am going to close the issue, as it is a Singularity version (or install?) issue.
I have tested on an other machine, with a different Singularity version and it works, without segmentation fault:

$ singularity pull --name star-2.6.1d--0.img https://depot.galaxyproject.org/singularity/star:2.6.1d--0
$ singularity exec star-2.6.1d--0.img STAR --version
STAR_2.6.1d

[vdanjean]

After investigation, the bug is due to the fact that the image (and STAR) is compiled with an old libc and the host where singularity was running has a new kernel without vsyscall support. Rebooting the host with the "vsyscall=emulate" solves the issue.

[rafalcode]

Luckily I looked in here, closed issues, because this was my problem and the vsyscall=emulate as-kernel-parameter solved it. Wouldn't be a bad idea to highlight this issue in some way, as it's pretty hidden in here.

[yxsee]

I am also getting this error when running STAR_salmon, but STAR_RSEM runs fine. Are there any plans on using a newer version of STAR, or a singularity image compiled without vsyscall?