Closed erinyoung closed 9 months ago
This is not yet enough information to go on. The mkfifo
failed points to one of the bash pipes failing, but I'm not certain yet. Can you paste the contents of /Volumes/IDGenomics_NAS/Bioinformatics/eriny/mosquito/2023-12-01/work/11/33708dd319d5e790a7d5b44e106865/.command.log
, please? And in the case they differ, also .command.out
and .command.err
? Thanks in advance.
Something else that you can try is, when in that work directory, run bash .command.run
and observe interactively what happens. You might also edit the .command.sh
file and include a line before the centrifuge command:
echo $db_name
to see if something went wrong there.
Here is the contents of .command.log
$ cat /Volumes/IDGenomics_NAS/Bioinformatics/eriny/mosquito/2023-12-01/work/11/33708dd319d5e790a7d5b44e106865/.command.log
INFO: Environment variable SINGULARITYENV_NXF_TASK_WORKDIR is set, but APPTAINERENV_NXF_TASK_WORKDIR is preferred
INFO: Environment variable SINGULARITYENV_NXF_DEBUG is set, but APPTAINERENV_NXF_DEBUG is preferred
(ERR): mkfifo(/tmp/72.inpipe1) failed.
Exiting now ...
The contents of .command.err look the same (to me) as .command.log
$ cat /Volumes/IDGenomics_NAS/Bioinformatics/eriny/mosquito/2023-12-01/work/11/33708dd319d5e790a7d5b44e106865/.command.err
INFO: Environment variable SINGULARITYENV_NXF_TASK_WORKDIR is set, but APPTAINERENV_NXF_TASK_WORKDIR is preferred
INFO: Environment variable SINGULARITYENV_NXF_DEBUG is set, but APPTAINERENV_NXF_DEBUG is preferred
(ERR): mkfifo(/tmp/72.inpipe1) failed.
Exiting now ...
The .command.out file is empty
$ cat /Volumes/IDGenomics_NAS/Bioinformatics/eriny/mosquito/2023-12-01/work/11/33708dd319d5e790a7d5b44e106865/.command.out
When I edited the .command.sh script to echo the $db_name variable, the result was nt_2018_3_3/nt
.
Is this the correct value?
I'm a bit at a loss as to the source of the error. @jfy133 do you have any ideas?
Actually, there could be a problem with your harddrive where /tmp
is mounted. You say the hardware is local, but is there anything special about it? Is it a network drive, for example? Or maybe it's simply running out of space? /tmp
on many linux systems is actually "mounted in memory", so it might be more limited than you think.
I get this sometimes too on my laptop during testing
My suspicion is that somehow there is some conflict when there are multiple centrifuge jobs running at the same time.
I think on my laptop I set maxForks to 1 and that helped reduce a lot those errors (IIRC).
It's also been reported on the centrifuge repo I think but no response
We have a slack thread with @erinyoung somewhere where I proposed it.
I'm getting this issue too. So far setting the centrifuge process to maxForks = 1 has reduced it, less than ideal solution though.
I'm getting this issue too. So far setting the centrifuge process to maxForks = 1 has reduced it, less than ideal solution though.
Spoke too soon.
So I believe indeed it's something where two centrifuge runs on the same node send up trying to make the same pipe name, as per: https://stackoverflow.com/questions/13040021/mkfifo-error-error-creating-the-named-pipe-file-exists
So I think if we can set a unique /tmp/ directory (as @LilyAnderssonLee tried before, but something went iffy), maybe we can get this to work
My solution doesn't work currently because the --temp-directory option isn't exposed ot the suer for some reason...
Perhaps centrifuge will honor the TMPDIR
environment variable? Then we could create a new temporary directory for each process invocation and set TMPDIR
to that.
Nope, it's literally hardcoded in the software: https://github.com/DaehwanKimLab/centrifuge/blob/9e244583481cc273f74415f2a2418e8fd342ab17/centrifuge#L357
I've openned an issue to ask: https://github.com/DaehwanKimLab/centrifuge/issues/268
Description of the bug
I downloaded the pre-built centrifuge database from https://benlangmead.github.io/aws-indexes/k2 (the full link is https://genome-idx.s3.amazonaws.com/centrifuge/nt_2018_3_3.tar.gz) to use with taxprofiler. It does not run as expected. I first brought this up in slack.
This is my database file (the listed kraken2 databases worked fine):
This is the error message:
Command used and terminal output
Relevant files
No response
System information
Nextflow version : 23.10.0 build 5889 Hardware : Local Executor : local Container engine: Singularity OS : CentOS Version of nf-core/taxprofiler : current