We wait for replication between tests and can check hashes as this point as well.

By adding these changes we will be able to fail earlier and isolate the test which produced the inconsistent data during replication.