Details
-
Bug
-
Status: Resolved
-
Urgent
-
Resolution: Duplicate
-
None
-
Correctness - Transient Incorrect Response
-
Critical
Description
Previously on digest mismatch we upgraded to CL.ALL, which harmed availability. However, because we do not have consistent range movements, when they are underway it is possible for both the coordinator of a query and one of its targets have stale ring information.
In this case, we would previously have detected a digest mismatch and ensured ALL nodes responded before answering, so that any single stale node would have been detected. In the new implementation, we try to get answers only from the nodes whose digests mismatched, so if the node were stale, it can respond with a stale data request also, reducing effective CL by one.
Attachments
Attachments
Issue Links
- duplicates
-
CASSANDRA-14740 BlockingReadRepair does not maintain monotonicity during range movements
- Resolved