Dowell-Lab/Nascent-Flow

Add Completion Tracking to Each Step

Closed this issue · 1 comments

zmaas commented

I'm working on the new NascentDB backend at the moment. One of the things we'd like to do is check whether or not a given step in the pipeline completed execution when importing NextFlow output into the database. This will require adding a few additional lines of script at the end of each step to write to a completion_status file, which could be stored alongside the software versions step. I'm working on an implementation now, with a pull request coming soon.

One key consideration here is that jobs might finish at the exact same moment as each other, so this code should probably do something like:

flock completion_status -c 'echo "fastqc: complete" > completion_status'

Using a file lock will make sure the output doesn't get clobbered if we have a race condition.

zmaas commented

This can effectively be performed using existing pipeline information. This is no longer necessary.