interrupted
Sounds extremely unlikely in my case. But what’s possible is that is that the backup job was started, and then upload failed due to TLS issue. The very first file of new backup version. That does sound likely indeed. But it shouldn’t trigger need for rebuild, just not-committing stuff that didn’t get through. Things that happen and are kind of unexpected make me curious. - So something illogical happened, which shouldn’t happen.
To be totally clear about this, I did try run test after the first backup failed. Just to find out if it’s backup or something else. Of course that test also failed with SChannel issues. Yet it shouldn’t change anything in the db.
version 2.0.4.34_canary_2019-11-05
No, rebuild consists several parts. The rebuild itself was really quick, I didn’t even notice when it happened. But with slow remote storage as example the JottaCloud, downloading files itself can take a really long time. And itself wasn’t bandwidth limited, but API latency limited. If you need to download 1000 5 kb files, it might take a long time, even if the amount of data getting transferred is small.