I was trying to verify my aes-key using commandline, got an error that the database did not exist, and noted it assumed a different name than the real one. The setup is still very simple, a single backup done (1 version). The backup table (1 row) in Duplicati-server.sqlite has the correct name.
I assume i can simply edit the dbconfig.json file. It looks to me that at some point an update of dbconfig.json was not done. All work except the verification attempt done from the gui.
If you have a specific DB you want on CommandLine, best plan is to say it.
GUI “Export as command line” can do that, or just type --dbpath yourself.
Editing dbconfig.json works too, but won’t stick on, say, destination change.
If you specify the
--dbpathparameter, it will not use thedbconfig.jsonfile and it will not store anything in the local datafolder.
It’s perfectly real, just separately chosen as needed – just as the GUI will do.
It’ not inherently “correct”. It’s just what GUI used, on Database screen too.
Not sure what this means. CommandLine is standalone, not tied to GUI plan.
This is by design. If you prefer a coordinated plan subject to one run at once,
ServerUtil can run (or queue) your GUI backup from CLI, if it’s of any interest.
Thanks for the replies.
I do not remember creating the json. And want to add, that creating another backup did not update the corrected version. I understand from the replies that is apparently is not meant to be. I urge you to increase the consistency between cli and gui.
That’s proper. You don’t create the json yourself. Duplicati creates it for you.
If you create another backup, it picks another database name automatically.
I might misunderstand, because some of the descriptions are not sufficient.
The whole point of this (which predates the GUI) is to manage DB locations.
The rest of the link that I gave explains the idea this way:
Using the CLI is simply a matter of invoking the binary:
C:\Program Files\Duplicati 2\Duplicati.CommandLine.exe help
Since the CLI also needs a local database for each backup, it will use the same location as described for the Server above to place databases. In addition to this, it will keep a small file called
dbconfig.jsonin the storage folder where it maps URLs to databases. The intention of this is to avoid manually specifying the--dbpathparameter on every invocation.
They’re consistently independent, but I didn’t design it. What’s the plan?
Remember that one can’t alter existing behavior without harming users.
Also recall that GUI can have many jobs, each having its own database.
CLI jobs have their own databases. Users can differ, preventing update.
I don’t even know what you mean by “consistency”, given such diversity.
Feel free to clarify, or (IMO better), use the system as it’s meant to work.
If you have a specific request, there is a Features category in the forum.
Putting it in GitHub is better because that tracks it, while forum does not.