I've tried a number of different sync programs and while SyncBack seems to be the best all round, it's unfortunately no good at all when it comes to handling file system errors. The current error handling system seems to just be "log it and move on" which actually isn't handling errors at all, it's just noting them down for a human to manually handle later.
I'm syncing to a windows network share over VPN so it's a regular windows network file copy, but over slow internet speeds and it occasionally generates transient "access denied" errors. I can't do anything about these. These errors always go away if you retry them, but SBSE does not do this. They are just logged and that's it, and if you're using the fast backup system (which is necessary due to network speeds) then on the next run, these failures go into the fast backup database as needing to be copied in full, which means lots of wasteful data copying - which can again fail for the same transient errors.
I can never get my backup to be 100% complete because of this, because scanning the destination will always result in a large number of random files and folders that failed with an error, and the only thing I can do is another full scan which results in another large number of random failures, or copy a large amount of data needlessly, some of which will also fail for the same errors and so a full scan is still needed to pick up the pieces after.
I've been using Free File Sync as well, which isn't particularly good software, but it does handle errors well - it has the option to retry any errors (both in analysis and in copy) X times with X second delay, and this solves all of my problems 100%. It is missing many other features that SBSE has, so I can't really use it. Yet it handles errors perfectly, and it's completely free! Currently I have to do a full scan and copy using Free File Sync after using SBSE, in order to be certain that I've actually achieved 100%.
SBSE would be absolutely fantastic if it could handle errors automatically - simply the option to "retry errors X times after X second delay" in the profile would be ideal, should be simple to implement, and it would make this genuinely useful and good software.
What would be even better is if any files that are still failed after X attempts during analysis, if you're running a fast backup, then these should go back into the fast backup database to be scanned again. They shouldn't go into the database to be copied in full, as they currently do - they have failed to be analysed, so on the next run they again need to again be analysed, not copied in full
Statistics: Posted by Statick — Wed Jun 10, 2020 1:36 pm