Closed prototaxites closed 4 months ago
@prototaxites would you be able to investigate this as a bug fix? I guess it could be quite small fix; or slightly more (but more robust) if it was converted to a proper nf-core module.
Can take a look but might take me some time to get to it, as my plate is quite full at the moment! I did build an updated version of the container back when I first made this issue, but I am not convinced I remember it fixing the issue.
Ultimately I think the best thing is to just replace the custom module with official nf-core ones... so hopefully it gets solved then (as will be using latest verison etc)
There are a bunch of issues which I think means this would be fixed by teh whole-sale update, closing in favour of the most recent issue as it's all related https://github.com/nf-core/mag/issues/611
Description of the bug
CAT_DB_GENERATE fails on my system due to what seem to be network connection issues, despite nodes having internet access:
Digging into the error a bit, this appears to be a recurring problem with some older biocontainers: https://github.com/bioconda/bioconda-recipes/issues/11583 - the CAT mulled container was built on 15/03/2021, before this issue was fixed. CAT image might want to be updated to use a newer base image, which hopefully avoids this problem.
In debugging this, I also noticed that CAT_DB_GENERATE only gets 1 cpu by default, but there's a
--nproc
option which specifies the number of cores available when building the Diamond database. Might be good to bump that up to speed up datebase building.Command used and terminal output
Relevant files
No response
System information
HPC, Slurm, Singularity, Linux, dev