As far as possible use the default synchronous override mode of replication. In synchronous override mode as long as the network is connected, the replication would be in synchronous mode and in cases of network failure, the updates can go to the Replicator Log without failing the updates.

If you need to use only the synchronous mode, then in cases when the network cannot be recovered, change the replication mode later to resolve the problem.

If an RVG has NTFS or ReFS mounted volumes and one of its associated Secondary hosts is in synchronous mode of replication, then in the case of a network failure all the writes to the replicated volume fail. The system may display a Delayed write failed error and even freeze or hang.