Write a test that kills the primary while a write is still uncommitted (using a failpoint) with a timestamp prior to a committed write. Ensure that when the node is restarted, the recovery process does not resurrect the committed write.
- depends on
-
SERVER-44555 Implement replicate before journaling
- Closed
- related to
-
SERVER-59150 non_durable_writes_on_primary_can_reach_majority.js does not wait for durable timestamp on secondary to advance
- Closed
-
SERVER-41392 Modify the _oplogJournalThreadLoop() to no longer call waitUntilDurable() and instead update the oplogTruncateAfterPoint
- Closed