Segmentation fault version 1.3.0

[FriederikeHanssen]

Hi,

I am running dragmap 1.3.0 on some small test data and continue to get a failure with

0 Segmentation fault      dragen-os -r dragmap --num-threads 2 -1 test_1.fastq.gz -2 test_2.fastq.gz 2> test.sorted.dragmap.log
          31 Done                    | samtools sort --threads 2 -o test.sorted.bam -

as soon as I downgrade back to 1.2.1, it works on the same data/setup.

Command run:

  dragen-os \
      -r dragmap \
       \
      --num-threads 2 \
      -1 test_1.fastq.gz -2 test_2.fastq.gz \
      2> test.sorted.dragmap.log \
      | samtools sort  --threads 2 -o test.sorted.bam -

[CarolineOhrman]

Hi, I am also experiencing the same issue with segmentationfault with version 1.3.0. Any progress?

My commands

reference indexing

dragen-os \
            --build-hash-table true  \
            --ht-reference input.fasta \
            --output-directory reference_folder/

mapping

dragen-os \
        -r reference_folder/ \
        -1 forward_pair.fastq.gz \
        -2 reverse_pair.fastq.gz \
        --num-thread 22| \
        samtools view -b  - | samtools sort -o output.bam 

[nservant]

same here :)

[micknudsen]

Has anybody managed to run v1.3.0 without getting a segmentation fault? If so, could you please share your conda environment using e.g. conda env export? I wonder if the problem is caused by one of the dependencies.

[ivanrwolf]

Same problem here

[gambalab]

same for me (installed from conda)

[VivekTodur]

Anyone figured how to get over this issue?

[edmundmiller]

Looks like https://github.com/Illumina/DRAGMAP/commit/c4cb6c94baaf6821db3fa696c53cdd809fb92304 might fix this, but the release was never cut on GitHub at least, so it hasn't triggered the conda builds and what not.

[ChengyueT]

I created a new env named dragmap and only installed conda dragmap in this environment, it worked for me.

[edmundmiller]

I created a new env named dragmap and only installed conda dragmap in this environment, it worked for me.

So maybe it's samtools? It also worked for @nvnieuwk 🤔

[edmundmiller]

BTW forked the repo because of various open PRs that didn't get the love they deserved, and merged them. Working on Rust FFI integration and going to start a rewrite if anyone's interested

https://github.com/edmundmiller/NARFMAP

[ghost]

Hi everyone. I still get the segmentation fault (core dumped) error message when using dragmap V1.3. It is installed with conda and also not working when installing it to a separate individual conda environment. Anyone has any advice what to do about it?

[ghost]

Hi everyone. I still get the segmentation fault (core dumped) error message when using dragmap V1.3. It is installed with conda and also not working when installing it to a separate individual conda environment. Anyone has any advice what to do about it?

It is working when I am using V1.2.1, so my input files should be fine.

[edmundmiller]

Yeah that's what we've found at nf-core 😞

You can try out https://github.com/edmundmiller/NARFMAP and I merged all of the open PRs #55 might have fixed it, but it didn't seem to.

[gambalab]

If you want you can try THIS singulairty version or the binary file in realease that should work on any debian distro.

[omixer]

If you want you can try THIS singulairty version or the binary file in realease that should work on any debian distro.

I can also confirm that the binaries from https://github.com/gambalab/DRAGMAP work well. I did few tests to validate my usecase. The dockerfile I used is below just in case.

FROM quay.io/biocontainers/dragmap:1.2.1--h72d16da_1

RUN wget https://github.com/gambalab/DRAGMAP/releases/download/1.3.1/dragen-os \
&& chmod 755 dragen-os \
&& mv dragen-os /usr/local/bin/

[brunorsci]

Não está estável ne na versão 1.2.1 nem na versão 1.3.1, ambas tem bugs.