Details
-
Sub-task
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
Description
Currently datanodes expect to receive delete transactions from SCM in ascending order. If they get a transaction less than the largest one they have seen, they will quietly drop it and respond with success to the SCM. In some corner cases it may be possible that transactions arrive out of order, so the datanodes should not depend on ordering to prevent a large backlog of duplicate delete transactions.
Attachments
Issue Links
- relates to
-
HDDS-8228 Add metric for out of order/retried delete transactions
-
- Resolved
-
- links to