Database Integrity for Backups with S3 Backend

Just a question about restoring backups when using S3 as the backend. Say I have my system set up to use S3 as the backend with encryption, and that I’m backing up my Seafile installation folder once per day. If I have a catastrophic hardware failure and my system dies halfway through the day, how does that affect my capacity to restore the system to its original state? Are all the database contents stored on S3 (and hence I will lose nothing)? Or is some filesystem data stored on the physical server which could result in some kind of mismatch between server and S3 data and loss/corruption of data? I know there is the fsck tool but I’m not sure how it interacts with S3 storage, if at all.


If your database backup is older than the data in S3, fsck will try to restore the libraries to the states of the database backup. So you’ll lose some recent data. But this is actually how backup works. You can setup a MySQL replication to reduce the window of data lose.