The fsync command on the secondary does not force the durable timestamp forward.
The test passes most of the time because the journal flusher updates the durable optime in the TopologyCoordinator before the test calls replSetGetStatus.
- is related to
-
SERVER-56522 Use proper timestamp comparison in non_durable_writes_on_primary_can_reach_majority.js
- Closed
-
SERVER-41386 Create a JS test that ensures non-durable writes can replicate but are not counted for majority write concern
- Closed
-
SERVER-41387 Create a test that ensures the oplogTruncateAfterPoint will correctly clear writes on startup after a primary restarts
- Closed