The application doesn't recognize/notify/handle gracefully if the two endpoints are out of sync

[legezam]

I noticed it happens when i switched from timeline based backups to manually triggered, number based approach on my server.

Steps to reproduce:

1. start out from a setup where the source uses timeline based snapshotting, have some snapshots be created
2. replicate those snapshots to the destination
3. disable timeline creation on source
4. cleanup (remove all) timeline snapshots from source
5. create a manual snapshot on source
6. run snapper-replicator

Snapper does the numbering based on the current maximum assigned number in a configuration instead of the maximum ever assigned number.

What happens

Since the timeline cleanup has thrown the previously assigned numbers away, snapper will reuse those numbers in the future. However, those numbers are already present in the destination so snapper-replicator silently ignores the new snapshots

What should happen

Ideally snapper-replicator should warn or let the user know about the inconsistency

Workaround

When doing the switch, before cleaning up the timeline snapshots, create a manual snapshot so a snapshot with high number remains in the configuration so snapper-replicator can continue the replication work in a consistent way