Testing the restore process

I am interested in running a periodic task to test the restore process as if there is no local database or source files (to simulate a restore in case of complete hard drive failure). Would the best approach be to run a restore with --no-local-db=true and --no-local-blocks=true? Any additional options that should be specified?

Also, would a restore with --dry-run=true --no-local-db=true be roughly equivalent to rebuilding the local database?

I believe you’ve got it right with this two parameters.

However I don’t think using --dry-run will simulate a DB restore (though I could be wrong, so let us know if you try it out).

Got it, thank you. How would I know if --dry-run with --no-local-db simulates a DB restore? I guess if it takes about as long as a real DB restore (about 15 minutes in my case), that would be a good indicator.

Probably. :slight_smile:

The tricky bit with --dry-run is certain things can’t be simulated with actually doing other things beforehand, so it can be unclear whether actual changes are being made in step 1 so the thing you want to test in step 2 can be simulated.

If you really want to see what a DB restore is like, it is safe to rename (or move) your current sqlite file and actually do a rebuild. You can then use the rebuilt one or delete it and put your original back in place. Note that you might want to disable the automatic scheduling of your job when doing this test.

Got it. Thanks for the tips.