Details
-
Bug
-
Resolution: Fixed
-
Major - P3
-
None
-
Fully Compatible
-
ALL
-
Sharding 2021-04-19
-
155
-
1
Description
The changes from c8e74f9 as part of SERVER-55723 were insufficient. An issue remains that this part of the future chain isn't being joined by the whenAll() / whenAllSettledOn() which comes later in the future chain. It is also possibe for chainCtx->oplogApplierConsistentButStaleFutures and chainCtx->oplogApplierStrictlyConsistentFutures to be reassigned while the whenAllSettledOn() are attempting to wait on them (read: data race).
Attachments
Issue Links
- is related to
-
SERVER-55723 ReshardingDataReplication may emplace _consistentButStale more than once
-
- Closed
-
- related to
-
SERVER-67466 Internal transactions API may be memory unsafe when outer OperationContext is interrupted
-
- Closed
-