Sure, this one is english version, let me know if this one is good enough:
What “non-authoritative recovery” means
You tell one DFSR member (A006) to discard its DFSR database for the replicated folder and rebuild it from a healthy partner. Files on disk stay in place, the member just relearns the replication metadata and treats the partner as the source of truth until it finishes the first sync.
When to use it
After power loss/dirty shutdown, or when health reports say the database/group is invalid and the member never leaves the exception/initializing state.
How to apply at a high level
Pick a healthy partner for A006. Put A006 into non-authoritative mode for that replicated folder so it rebuilds its DFSR DB from the partner. Let the initial sync complete; A006 will then return to normal bidirectional replication.
Important notes
Do this to one member at a time (A006), not to every server. Keep DNS/time/AD healthy and ensure staging has enough free space (at least larger than your biggest file). If recovery keeps failing or multiple members show “invalid,” recreating the replication group is often faster.
If you’d rather avoid recovery steps altogether:
Remove A006 from the replicated folder, preseed its data (to keep ACLs/timestamps), then add A006 back and let it converge with a single Primary member for the first sync.