Just in case I'm missing anything obvious: I'm trying to run this pipeline on some SmartSeq2 data and ran into this issue github.com/nf-core/rnaseq/issues/1912 possibly related to https://github.com/nextflow-io/nextflow/issues/7247 (Apologies for cross-posting)

Sorry, seems pretty clearly a file system thing, outside the control of this workflow, so I'm closing the issue here. You could comment on that issue you linked to highlight your failure, and check on the ongoing work to see if you think it would help: github.com/nextflow-io/nextflow/pull/7281

It may be worth reposting the answer on #C043FMKUNLB but since I got you here, I'm a little confused by the description of ``executor.exitReadTimeout`` docs.seqera.io/nextflow/reference/config/executor#… I've ramped it up to 30 min like I saw in other Slurm

nf-core/configs

and the pipeline seems to be running fine now (in progress). Clearly, jobs are submitted long before 30 min after previous jobs complete, so just to be sure I understand: •

exitReadTimeout

doesn't apply to the time between jobs that follow each other? • subsequent jobs are submitted as soon as upstream jobs are completed, without necessarily having seen the

.exitcode

file? • the pipeline might fail

exitReadTimeout

time after a job completed with an non 0 exit code?

Yep, you really should ask questions like this on the more generic channels, or on the Nextflow Slack. You'll get snappier, more useful responses! But... The job does need to see the exitcode to be considered done. So if, in a tricky file system, it takes 20 minutes for that file to appear, it WILL add a 20 minute delay before the downstream task starts. But

exitReadTimeout

is a ceiling you shouldn't be getting near in most cases, so it won't add those delays unless there's something very wrong with the infrastructure which causes those delays to happen regularly.