You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Requirement originated from code review comment related to "batch completed" notification processing to generate manifests: #214 (comment)
Original comment:
Are there any potential issues with this if we scaled up? E.g. if there was a Manifest that had 3 batches and we had scaled to 3 engines and 3 "complete" listeners. If each of those listeners received a "complete" notification at the same time for each of the 3 batches - is there a chance something could fall through, and the 3 batches are separately marked as 'complete' at the same time so the final "done" logic isn't fired?
I wonder if we could add a optimistic concurrency token to Batch to handle that? ie one of the 3 batches would fail and be requeued (not for this ticket, can revisit)
Ultimately the requirement is to prevent concurrent competing notification all marking separate batches as 'done' but no 1 single processor running the "batch completion" logic, meaning the Manifest wouldn't be generated.
Requirement originated from code review comment related to "batch completed" notification processing to generate manifests: #214 (comment)
Original comment:
Ultimately the requirement is to prevent concurrent competing notification all marking separate batches as 'done' but no 1 single processor running the "batch completion" logic, meaning the Manifest wouldn't be generated.
An optimistic concurrency token was suggested as EF has native support for this: https://www.npgsql.org/efcore/modeling/concurrency.html?tabs=fluent-api and it would be a relatively rare occurrence.
The text was updated successfully, but these errors were encountered: