Changing the reference genome in minigraph_cactus run results in failed job - Job used more disk than requested

[evcurran]

Hello,

I'm running minigraph cactus on a single chromosome across 7 haplotypes as a test run, before scaling up to the whole genome (8 chromosomes, ~ 200Mb), for around 24 haplotypes. Downstream I plan on using vg giraffe for SV genotyping short read data.

The test run successfully completed, with odgi visualisations. The haplotype I used as the reference for this run was the smallest, so I swapped it out for the largest haplotype (changing no other settings), and then it failed before completion, due to memory related issues and a lack of requested disk space.

I found it unclear how to increase requested disk space, is this done in the config.xml file? Or are there some other underlying issues here? Since I plan on scaling this up, it would be good to get some clarity on how to specify the appropriate settings.

I have attached the log file with the error and the job submission script with the command:

minigraph_cactus_chr1_filt2_velH1.err.txt

minigraph_cactus_chr1_test_velH1ref.sh

Thank you for any help with this!
Emma

[glennhickey]

The disk space thing is confusing, but nothing to worry about. The issue is the odgi layout is crashing. I'm not sure what's going on with it. My suggestion is to run without odgi, which should work. Then try to run the latest odgi version yourself to make any visualizations from the gfa and post any problems on the odgi github...