Configuring Hyper-V Replica Retry Policy to Prevent Duplicate Write Operations

The goal is to configure Hyper‑V Replica so that a retry of a failed replication pass does not cause the same write to be applied twice on the replica VM.

When a replication pass encounters a transient network error, Hyper‑V may retry the transfer of the changed blocks. If the retry resends blocks that were already partially committed, the replica could see duplicate writes unless the underlying write ordering or checkpoint mechanism prevents it.

What retry interval and maximum attempt values, combined with the replica’s write‑log checkpointing behavior, guarantee that retries are safe without duplicating writes?