What should I check when cloud data has been synchronised across devices?¶
Treat synchronisation as a series of separate cloud and device events, not as one magical state where every copy is instantly identical.
Different devices may hold different versions, timestamps, caches or deletion states.
Identify every relevant copy¶
Record, where available:
- cloud resource ID;
- version ID;
- account;
- device;
- sync application;
- local path;
- local file identifier;
- provider timestamps;
- local filesystem timestamps; and
- sync status.
A simple three-copy picture might look like this:
The tablet may later reconnect and create another event from an older copy.
Synchronisation can produce several kinds of local artefact¶
A device may contain:
- full local file;
- online-only placeholder;
- preview;
- cache;
- conflict copy;
- deleted local copy;
- local version awaiting upload; or
- database entry describing sync state.
Do not assume the visible filename tells you which of these you have.
Preserve conflict and offline states¶
Conflict copies can be particularly useful because they show that two versions existed independently.
A disconnected device may retain content that has since been deleted in the cloud.
Avoid reconnecting an isolated device until the likely effect is understood. Synchronisation may upload, overwrite, rename or delete material and create new provider events.
How do I identify which device was linked to cloud activity? helps when the provider record needs to be joined to hardware.
The practical point is: cloud sync is a timeline of copies and transfers. Preserve the differences rather than forcing every timestamp and version into one story.