Command terminated by signal 8

[clemgoub]

Hello pggb team!

I've been able to run pggb successfuly on a couple of genomes (~150Mbp each, eukaryotes), and ramped up the analysis to 14 genomes. However, I encountered a new error (signal 8) after the pipeline produces the GFA:

I am using pggb from the main git branch, b2843d0fb7a12c0103427f30c55905ba0565ad41 I built a docker image from the Dockerfile, using the method to build locally.

Then:

docker run -it -v /home/cgoubert/projects/Fly/data/genomes/d_melanogaster/Rech2022/pggb:/data cgoubert/pggb /bin/bash -c "pggb -i /data/dm6.plus.Rech13_PanSN-spec.fa.gz -n 14 -p 90 -s 5000 -V dm6:200 -o /data/102423_pggb_dm6.Rech13 -t 32"

[...]
[seqwish::transclosure] 566.773 100.00% building node_iitree and path_iitree indexes
[seqwish::transclosure] 566.775 100.00% done
[seqwish::transclosure] 566.795 done with transitive closures
[seqwish::compact] 566.795 compacting nodes
[seqwish::compact] 568.448 done compacting
[seqwish::compact] 568.644 built node index
[seqwish::links] 568.644 finding graph links
[seqwish::links] 568.649 links derived
[seqwish::gfa] 568.649 writing graph
Command terminated by signal 8

The output folder shows:

-rw-r--r-- 1 cgoubert repeat 1.7K Oct 24 11:56 dm6.plus.Rech13_PanSN-spec.fa.gz.0ec1561.417fcdf.8abe853.smooth.10-24-2023_18:56:13.log
-rw-r--r-- 1 cgoubert repeat 1.5K Oct 24 11:56 dm6.plus.Rech13_PanSN-spec.fa.gz.0ec1561.417fcdf.8abe853.smooth.10-24-2023_18:56:13.params.yml
-rw-r--r-- 1 cgoubert repeat 5.2M Oct 24 20:15 dm6.plus.Rech13_PanSN-spec.fa.gz.0ec1561.417fcdf.8abe853.smooth.10-24-2023_18:59:58.log
-rw-r--r-- 1 cgoubert repeat 1.5K Oct 24 11:59 dm6.plus.Rech13_PanSN-spec.fa.gz.0ec1561.417fcdf.8abe853.smooth.10-24-2023_18:59:58.params.yml
-rw-r--r-- 1 cgoubert repeat 1.9G Oct 24 20:15 dm6.plus.Rech13_PanSN-spec.fa.gz.0ec1561.417fcdf.seqwish.gfa
-rw-r--r-- 1 cgoubert repeat 525M Oct 24 20:05 dm6.plus.Rech13_PanSN-spec.fa.gz.0ec1561.alignments.wfmash.paf
-rw-r--r-- 1 cgoubert repeat 2.5M Oct 24 12:03 dm6.plus.Rech13_PanSN-spec.fa.gz.0ec1561.mappings.wfmash.paf
drwx------ 2 cgoubert repeat  164 Oct 24 20:15 seqwish-3xM1GS

Thanks for your help!

Clément

[AndreaGuarracino]

Can you try again, but specifying a temporary folder where you are sure to have lots of space and that is a fast SSD? Usually, it is -D /scratch on clusters

[clemgoub]

Dear @AndreaGuarracino, sorry I can only reply now. The server I'm currently using is a single node without a dedicated scratch disk. We have plenty of room for storage however on the disk I used for outputs. Would there still be some workaround/testing I could implement?

I am curious to know why it returns a signal 8 (floating point exception) if we have a disk issue? Would it be indicating that it is taking too long to write?

Thanks for you help!

Clém