I updated to version 2.1.0.2 yesterday. After that, I get this error:
The operation Backup has failed with error: The bucket you are attempting to access must be addressed using the specified endpoint. Please send all future requests to this endpoint.
AmazonS3Exception: The bucket you are attempting to access must be addressed using the specified endpoint. Please send all future requests to this endpoint.
I’m wondering whether something has been changed in the version regarding the S3 connectivity.
The S3 backend has been updated to the latest library version, but no additional changes are applied.
This usually indicates that you are using the general S3 endpoint, but must use a region specific endpoint. If you are using the UI, you should be able to pick the region endpoint in the dropdown.
Yes, I’m using the UI and I can pick the right region.
Meanwhile I tried to test the connection with the CLI:
duplicati-cli test s3://mrinfosite.de.backup/develop --aws-access-key-id=<myId> --aws-secret-access-key=<mykey> --s3-server-name=s3.eu-west-1.amazonaws.com --s3-location-constraint=eu-west-1 --log-file=/tmp/duplicati_log.txt --log-file-log-level=Verbose
This show the error: The operation Test has failed with error: Database file does not exist: /home/user/.config/Duplicati/CKRVRHGNZZ.sqlite
In file /home/user/.config/Duplicati/ZNZMRGHYKE.sqlite is available.
Maybe it’s important to mention that I removed the Duplicati package before I installed the new version.
Can this cause the issue?
Do you usually use CLI? If you usually use GUI, a compatible CLI test needs --dbpath=<DB>.
The TEST command explains how test is backup-specific. That’s why it needs its database.
Not giving one means it has to assign one. It gets associated with the target URL you gave it.
BackendTool and BackendTester are more pure communication tests. You can get a URL with Export As Command-lineof a GUI job, however I’m not certain if all options attach to the URL.
Okay, understood. I went back to the GUI. And did a database repair. This fixes my problem with S3 data point. But it looks like that the backup starts to copy all files again, instead of doing just the increment.
I don’t know how database repair could fix communications, but nice…
If previous backup was not successfully finished (I lose track), some amount of re-reading may be necessary. It also doesn’t remember its position, so you’ll at least see a lot of looking for changes.
What you shouldn’t see is lots of upload. You can watch About → Show log → Live → Information
In fact I did two steps: First I entered the correct region to fix the communication to S3. Second I recreated the database. So I’m happy now.
Thank you for help.
The CLI and the Server (UI) uses two different ways of keeping track of the local database. The UI uses the file ~/.config/Duplicati/Duplicati-server.sqlite to store the full path to the database for each configured backup. You can use the menu “Database …” to see (and copy) the path.
The CLI uses the file ~/.config/Duplicati/dbconfig.json to keep track of auto-generated databases. You can use the option --dbpath=<path to database> to bypass the dbconfig.json file as @ts678 mentions.
The error message indicates that the CLI did at some point create an entry in dbconfig.json but that database is no longer there. If you grab the path from the UI, you can edit dbconfig.json or use --dbpath to get the CLI working on the same database.