Unfortunately, the --no-backend-verification
just disables the step that is throwing the errors, it doesn’t actually do anything to resolve it.
You are correct, a database repair alone should resolve such “files that are missing from the remote storage
” type issues. I’m not really sure why it didn’t for you - and I know a few other people have reported this happening as well.
My guess is either:
-
The real error is something else but getting caught / reported as a “missing files” problem (thus a repair / recreate isn’t fixing something that isn’t broken)
-
There’s another issue on top of the “missing files” problem that causes the repair / recreate to not work correctly.
-
There’s an edge case bug in the repair / recreate process that isn’t being properly identified and handled.
I ran into this situation last month when doing some simulated failure recoveries. In my case a Repair (not recreate) was all that was needed to resolve that issue, though I did suggest a warning resulting from failing to delete a file that couldn’t be found is OK to show ONCE but the file should then be considered deleted.
If @ksignorini is willing to test, it would be interesting to find out if this works. I’m torn on whether it will resolve the issue or move to a different set of warnings about the file sizes or dates being wrong.