Open sfayer opened 5 years ago
looks like the db is being updated with a value that fails the constraint, specifically the attempts constraint.
Where unless specified max_tries defaults to 2: https://github.com/ic-hep/pdm/blob/a98949a6717be2d84e1599918a3a89d94741fb9a/src/pdm/workqueue/WorkqueueDB.py#L298
This could be a race condition... However given we don't see this with the BY_NUMBER
algorithm seems that this is to do with the BY_SIZE
algorithm. This has been changed extensively in #436 so maybe we should wait and see if this is still an issue after that PR is merged
We got this stack trace which testing the BY_SIZE algorithm (possibly multiple worker instances updating things at the same time or something like that)?