Command terminated by signal 9

[ChuanzhengWei]

Hi,

I encountered an error during the process, specifically 'Command terminated by signal 9'. I'm unsure of the cause, as 7 out of 10 of the community runs were successful, and only 3 failed. Could it be related to the sequence assembly?

The files generated by the unsuccessful communities are as follows.

The location of the error in the log.

[2025-02-12T05:51:35Z INFO  gfaffix] running gfaffix on 1 threads
[2025-02-12T05:51:35Z INFO  gfaffix] loading graph from /data/community3/CNSb.community.3.fa.gz.35d2267.11fba48.f50f1bb.smooth.gfa
[2025-02-12T05:51:38Z INFO  gfaffix] constructing handle graph
[2025-02-12T05:51:39Z INFO  gfaffix] handlegraph has 1522912 nodes and 2091587 edges
[2025-02-12T05:51:39Z INFO  gfaffix] storing length of original nodes for bookkeeping
[2025-02-12T05:51:39Z INFO  gfaffix] identifying walk-preserving shared affixes
[2025-02-12T06:35:37Z INFO  gfaffix] node 282461 is a palindrome
[2025-02-12T06:35:37Z INFO  gfaffix] node 282461 is a palindrome
Command terminated by signal 9
gfaffix /data/community3/CN.community.3.fa.gz.35d2267.11fba48.f50f1bb.smooth.gfa -o /data/community3/CN.community.3.fa.gz.35d2267.11fba48.f50f1bb.smooth.fix.gfa
6560.16s user 200.80s system 99% cpu 6769.34s total 142598068Kb max memory
pigz
0.00s user 0.03s system 0% cpu 6769.98s total 908Kb max memory

I would appreciate any insights or suggestions you might have on this issue.
Thank you for your help, and I look forward to your response.

[AndreaGuarracino]

How much RAM does your machine have? It seems gfaffix got hungry and went out of memory. Can you report the gfaffix version? It might be a bug there or just a graph too complex to normalize, requiring a machine with more RAM. @danydoerr

[ChuanzhengWei]

singularity run -B ${PWD}/data:/data pggb.sif pggb --version
pggb v0.7.2
singularity run -B ${PWD}/data:/data pggb.sif gfaffix --version
gfaffix 0.2.0

[ChuanzhengWei]

I've noticed that the failed tasks consume a significant amount of memory and processing time, while the successful tasks complete very quickly. I'm not sure what is causing this discrepancy.

Below are the logs for a successful run and a failed run.

**# successful**
[2025-02-11T15:24:48Z INFO  gfaffix] done
gfaffix /data/community7/xx.community.7.fa.gz.35d2267.11fba48.f50f1bb.smooth.gfa -o /data/community7/xx.community.7.fa.gz.35d2267.11fba48.f50f1bb.smooth.fix.gfa
5.55s user 0.76s system 95% cpu 6.61s total 584668Kb max memory
[2025-02-11T15:10:03Z INFO  gfaffix] done
gfaffix /data/community8/xx.community.8.fa.gz.35d2267.11fba48.f50f1bb.smooth.gfa -o /data/community8/xx.community.8.fa.gz.35d2267.11fba48.f50f1bb.smooth.fix.gfa
11.33s user 1.04s system 95% cpu 13.02s total 634488Kb max memory

**# failed**
[2025-02-12T14:55:45Z INFO  gfaffix] running gfaffix on 1 threads
[2025-02-12T14:55:45Z INFO  gfaffix] loading graph from /data/community9/xx.community.9.fa.gz.35d2267.11fba48.f50f1bb.smooth.gfa
[2025-02-12T14:55:47Z INFO  gfaffix] constructing handle graph
[2025-02-12T14:55:48Z INFO  gfaffix] handlegraph has 1302948 nodes and 1804496 edges
[2025-02-12T14:55:48Z INFO  gfaffix] storing length of original nodes for bookkeeping
[2025-02-12T14:55:48Z INFO  gfaffix] identifying walk-preserving shared affixes
Command terminated by signal 9
gfaffix /data/community9/xx.community.9.fa.gz.35d2267.11fba48.f50f1bb.smooth.gfa -o /data/community9/xx.community.9.fa.gz.35d2267.11fba48.f50f1bb.smooth.fix.gfa
27203.66s user 947.46s system 99% cpu 28209.42s total 481029964Kb max memory

Command terminated by signal 9
gfaffix /data/community2/xx.community.2.fa.gz.35d2267.11fba48.f50f1bb.smooth.gfa -o /data/community2/xx.community.2.fa.gz.35d2267.11fba48.f50f1bb.smooth.fix.gfa
26702.10s user 830.46s system 99% cpu 27547.65s total 481076052Kb max memory

[danydoerr]

@ChuanzhengWei I'm working on a much memory efficient version, can't say when it's ready to release. If you want to have a prerelease, please let me know.

[ChuanzhengWei]

I'm definitely interested in participating in the prerelease testing. Please let me know if there are any steps I should follow or further details you'd like to share. @danydoerr

In addition, could you please help clarify my question? I'm uncertain whether the error in my task is due to an issue with gfaffix or with my sequence assembly. I’ve noticed that while pggb runs successfully on some chromosomes, it fails on others.

[danydoerr]

@ChuanzhengWei I'm not an expert on pggb. GFAffix has memory issues on pangenome graphs that contain a lot of paths, and this is what I'm concerned about/trying to get fixed :)

[ChuanzhengWei]

Thank you for your response!

The error I encountered occurred at the GFAffix step,#447 (comment) but I’m not sure why. Do you think your prerelease version might resolve this issue? I’d love to give it a try if possible.

[danydoerr]

@ChuanzhengWei I'm also not sure why your job was killed, but insufficient memory could be the cause. Here's a pre-release of gfaffix that would solve this issue. Please note that this version does not accept graphs with W-lines. So, as long as you're working with pggb graphs, this should be fine.

[ChuanzhengWei]

I'm sorry to tell you that the pre-release version of gfaffix produced the same error. However, I then installed pggb using conda, and to my surprise, it ran successfully. The gfaffix version I used was 0.1.5b. Although I'm not entirely sure what caused the error, I'm fortunate that I was able to complete my work. If you're curious about the issue, I'd be happy to share my input files with you.
Thank you very much for your help. @danydoerr

[danydoerr]

@ChuanzhengWei Yes, I'd love to have your input files to get to the bottom of this issue. Could you please send me an email (daniel DOT doerr AT hhu DOT de) with instructions on how I can download them? Thanks.