After SERVER-125964 there is a new parameter initialSyncWaitForSyncSourceLastStableRecoveryTs which causes the node to wait for its sync source lastStableRecoveryTimestamp to reach beginApplyingTimestamp. In this test on slow variants we can hit a scenario where writes on our sync source were not majority committed, but our beginApplyingTimestamp is stuck ahead of lastStableRecoveryTimestamp permanently since we have disconnected the majority of the replica set.
- is related to
-
SERVER-135079 [v9.0] Revert SERVER-128583 Fix flakiness in resync_majority_member.js
-
- Closed
-
- related to
-
-
- Closed
-