Open e-t-k opened 8 years ago
Part 1 is fixed in https://github.com/UCSC-MedBook/MedBook-JobRunner/commit/7ebb22e00b686181a28be1256112ec36ab499ad2 and now live on prod ; part 2 ( if we're waiting on prerequisite, retry indefinitely) hasn't been added.
https://github.com/UCSC-MedBook/MedBook-JobRunner/blob/master/webapp/server/startup.js#L73
If a submission has any prerequisite jobs in any status (except error) it never checks if they are done, just forces it to retry for 2 minutes, then kills it. (if they're in error, it fails the job.)
So 2 things to change:
(Right now, the latter should technically never be a problem, since it won't try to pick up the new job until all other jobs are done; but this might change in the future & its an easy fix.)