Details
-
Bug
-
Resolution: Fixed
-
Major
-
None
-
Triaged
-
1
-
No
Description
If we de-dupe in the CheckpointManager then we don't update dirtyQueueAge accordingly but we do attempt to remove the original queueAge from dirtyQueueAge at flush.
We should also make VBucket::dirtyQueueAge a NonNegativeCounter to simplify a bunch of code and ensure it never goes negative.
Attachments
Issue Links
- has to be done after
-
MB-43373 Exceptions thrown in FollyExecutorPool tasks can hang KV-Engine tasks
- Closed