We need to validate the write durability guarantees of MongoDB under the scenario where the primary acknowledges a majority write operation, but the corresponding oplog entries are not yet durable due to the Secondary Log Server (SLS) failure or a crash/power loss on the primary node itself. In this situation, there is a risk that the acknowledged data becomes irretrievably lost.

Using Antithesis, we should simulate node failures and environmental disruptions (e.g., power loss, crash) during critical operations to assess the behavior of MongoDB and confirm whether the failure results in data inconsistency or loss.