Release: 2.1.0.111 (Canary) 2025-03-15

I was mostly poking to get an idea of how to reproduce this.

So the setup is:

  • Use 2.1.0.110
  • Choose a password
  • Use 2.1.0.111 (or 2.1.0.112)
  • Try to log in, and it fails?

But since you are using Docker, could it be that you get different Duplicati-server.sqlite files for each version?

no. i used the same. i also installed a clean version (2.1.0.111) and had the same issue there.

Not a new database, I just upgraded versions.
Yes, the only backup configuration I had on that PC was already in there.

Any idea about not being able to Continue forward on the first page of the edit configuration?

Thanks for the update!

So, I did some further tests.

  1. I deleted the content of my data directory, means I started with a clean install
  2. Installed 2.1.0.112
  3. Login not possible
  4. Installed 2.1.0.5
  5. Logon ok
  6. Installed 2.1.0.111
  7. Logon not possible
  8. Installed 2.1.0.110
  9. Logon ok

Used the same docker settings (Duplicati data is persisted in /nas/Podman/Duplicati):

-v /nas/Podman:/data
...
-e SETTINGS_ENCRYPTION_KEY=changeme \
-e DUPLICATI__WEBSERVICE_PASSWORD=changeMe!12 \

It seems, there is an issue from 2.1.0.111 onwards with the ā€œDUPLICATI__WEBSERVICE_PASSWORDā€ in docker.

EDIT:
Unfortunately, I cannot find any valuable logfile. journalctl shows only the start entries from podman and there I don’t see any differences between the versions.

I have only the browser log from a failed login.

1 Like

Same login fail (docker). Works if rollback or go stable.

No it still failed and also in .112 but a different flavour of Linux: Release: 2.1.0.112 (Canary) 2025-03-26 - #4 by Taomyn

I think I missed this part in my previous response. I think this is on the ngclient UI? (The one with a D icon top-left).

If so, I can reproduce this only if the name is empty. If you do have a name, try editing it and see if the button lights up. If not, you can also click the navigation dots above to jump to the page you want.

If it still fails, you can click ā€œRevert to NGAX clientā€ in the menu on the left to get the previous UI. But I would like to fix the issue. Do you have additional details on this?

Thanks to all for reporting this.

I have tracked it down to a version problem:

2 Likes

Minor issue with remote ngclient…

  1. Open hostname:8200, get known issue of blank screen
  2. Change URL to ngax
  3. GUI prompts for password
  4. Enter correct password
  5. GUI reverts to ngclient

Having switched the URL to ngax and having gone through the password dialog, I would expect the GUI to stay with ngax rather than reverting to ngclient.

That is controlled by the ā€œthemeā€.
You can go to hostname:8200/theme.html and choose the client.
This will set a cookie with the preferred client, and this client will then stick for 90 days.

The buttons to switch between the clients do not set this cookie.
In my view the switch buttons are temporary and do not indicate a general desire to use one client over the other. Let me know if that assumption seems wrong.

Like I said, minor issue.

But the end-user is going to see the interaction as being:

  1. temporarily switch to old UI
  2. ā€œinterruptā€ - please provide password
  3. enter password
  4. return from ā€œinterruptā€, but not to where we were before the ā€œinterruptā€

And I whole-heartedly agree that there are more important issues to address!

1 Like